Configuration reference
The agentgateway schema is available as a JSON schema. Review this page for more information about the schema and how to use it.
Config file validation
Many integrated development environments (IDEs) and editors support schema validation for your standalone agentgateway configuration file.
Default schema validation off main
The examples throughout the docs use the following schema that redirects to the agentgateway config on main.
# yaml-language-server: $schema=https://agentgateway.dev/schema/configVersion-specific schema validation
Replace $VERSION in the following schema with the version of agentgateway that you are using, such as 0.12.0.
# yaml-language-server: $schema=https://raw.githubusercontent.com/agentgateway/agentgateway/refs/tags/$VERSION/schema/config.jsonFor example:
# yaml-language-server: $schema=https://raw.githubusercontent.com/agentgateway/agentgateway/refs/tags/v0.12.0/schema/config.jsonSchema
The following table shows the complete agentgateway configuration file schema, with columns for the field and description.
| Field | Description |
|---|---|
config |
|
config.enableIpv6 |
|
config.localXdsPath |
Local XDS path. If not specified, the current configuration file will be used. |
config.caAddress |
|
config.caAuthToken |
|
config.xdsAddress |
|
config.xdsAuthToken |
|
config.namespace |
|
config.gateway |
|
config.trustDomain |
|
config.serviceAccount |
|
config.clusterId |
|
config.network |
|
config.adminAddr |
Admin UI address in the format “ip:port” |
config.statsAddr |
Stats/metrics server address in the format “ip:port” |
config.readinessAddr |
Readiness probe server address in the format “ip:port” |
config.session |
Configuration for stateful session management |
config.session.key |
The signing key to be used. If not set, sessions will not be encrypted. For example, generated via openssl rand -hex 32. |
config.connectionTerminationDeadline |
|
config.connectionMinTerminationDeadline |
|
config.workerThreads |
|
config.tracing |
|
config.tracing.otlpEndpoint |
|
config.tracing.headers |
|
config.tracing.otlpProtocol |
|
config.tracing.fields |
|
config.tracing.fields.remove |
|
config.tracing.fields.add |
|
config.tracing.randomSampling |
Expression to determine the amount of random sampling. Random sampling will initiate a new trace span if the incoming request does not have a trace already. This should evaluate to either a float between 0.0-1.0 (0-100%) or true/false. This defaults to ‘false’. |
config.tracing.clientSampling |
Expression to determine the amount of client sampling. Client sampling determines whether to initiate a new trace span if the incoming request does have a trace already. This should evaluate to either a float between 0.0-1.0 (0-100%) or true/false. This defaults to ’true'. |
config.tracing.path |
OTLP path. Default is /v1/traces |
config.logging |
|
config.logging.filter |
|
config.logging.fields |
|
config.logging.fields.remove |
|
config.logging.fields.add |
|
config.logging.level |
|
config.logging.format |
|
config.metrics |
|
config.metrics.remove |
|
config.metrics.fields |
|
config.metrics.fields.add |
|
config.backend |
|
config.backend.keepalives |
|
config.backend.keepalives.enabled |
|
config.backend.keepalives.time |
|
config.backend.keepalives.interval |
|
config.backend.keepalives.retries |
|
config.backend.connectTimeout |
|
config.backend.poolIdleTimeout |
The maximum duration to keep an idle connection alive. |
config.backend.poolMaxSize |
The maximum number of connections allowed in the pool, per hostname. If set, this will limit the total number of connections kept alive to any given host. Note: excess connections will still be created, they will just not remain idle. If unset, there is no limit |
config.hbone |
|
config.hbone.windowSize |
|
config.hbone.connectionWindowSize |
|
config.hbone.frameSize |
|
config.hbone.poolMaxStreamsPerConn |
|
config.hbone.poolUnusedReleaseTimeout |
|
binds |
|
binds[].port |
|
binds[].listeners |
|
binds[].listeners[].name |
|
binds[].listeners[].namespace |
|
binds[].listeners[].hostname |
Can be a wildcard |
binds[].listeners[].protocol |
|
binds[].listeners[].tls |
|
binds[].listeners[].tls.cert |
|
binds[].listeners[].tls.key |
|
binds[].listeners[].tls.root |
|
binds[].listeners[].tls.cipherSuites |
Optional cipher suite allowlist (order is preserved). |
binds[].listeners[].tls.minTLSVersion |
Minimum supported TLS version (only TLS 1.2 and 1.3 are supported). |
binds[].listeners[].tls.maxTLSVersion |
Maximum supported TLS version (only TLS 1.2 and 1.3 are supported). |
binds[].listeners[].routes |
|
binds[].listeners[].routes[].name |
|
binds[].listeners[].routes[].namespace |
|
binds[].listeners[].routes[].ruleName |
|
binds[].listeners[].routes[].hostnames |
Can be a wildcard |
binds[].listeners[].routes[].matches |
|
binds[].listeners[].routes[].matches[].headers |
|
binds[].listeners[].routes[].matches[].headers[].name |
|
binds[].listeners[].routes[].matches[].headers[].value |
|
binds[].listeners[].routes[].matches[].headers[].value.(1)exact |
|
binds[].listeners[].routes[].matches[].headers[].value.(1)regex |
|
binds[].listeners[].routes[].matches[].path |
|
binds[].listeners[].routes[].matches[].path.(1)exact |
|
binds[].listeners[].routes[].matches[].path.(1)pathPrefix |
|
binds[].listeners[].routes[].matches[].path.(1)regex |
|
binds[].listeners[].routes[].matches[].method |
|
binds[].listeners[].routes[].matches[].query |
|
binds[].listeners[].routes[].matches[].query[].name |
|
binds[].listeners[].routes[].matches[].query[].value |
|
binds[].listeners[].routes[].matches[].query[].value.(1)exact |
|
binds[].listeners[].routes[].matches[].query[].value.(1)regex |
|
binds[].listeners[].routes[].policies |
|
binds[].listeners[].routes[].policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].policies.requestRedirect.authority |
|
binds[].listeners[].routes[].policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].policies.requestRedirect.path |
|
binds[].listeners[].routes[].policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].policies.requestRedirect.status |
|
binds[].listeners[].routes[].policies.urlRewrite |
Modify the URL path or authority. |
binds[].listeners[].routes[].policies.urlRewrite.authority |
|
binds[].listeners[].routes[].policies.urlRewrite.authority.(any)(1)full |
|
binds[].listeners[].routes[].policies.urlRewrite.authority.(any)(1)host |
|
binds[].listeners[].routes[].policies.urlRewrite.authority.(any)(1)port |
|
binds[].listeners[].routes[].policies.urlRewrite.path |
|
binds[].listeners[].routes[].policies.urlRewrite.path.(any)(1)full |
|
binds[].listeners[].routes[].policies.urlRewrite.path.(any)(1)prefix |
|
binds[].listeners[].routes[].policies.requestMirror |
Mirror incoming requests to another destination. |
binds[].listeners[].routes[].policies.requestMirror.backend |
|
binds[].listeners[].routes[].policies.requestMirror.backend.(1)service |
|
binds[].listeners[].routes[].policies.requestMirror.backend.(1)service.name |
|
binds[].listeners[].routes[].policies.requestMirror.backend.(1)service.name.namespace |
|
binds[].listeners[].routes[].policies.requestMirror.backend.(1)service.name.hostname |
|
binds[].listeners[].routes[].policies.requestMirror.backend.(1)service.port |
|
binds[].listeners[].routes[].policies.requestMirror.backend.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].policies.requestMirror.backend.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].policies.requestMirror.percentage |
|
binds[].listeners[].routes[].policies.directResponse |
Directly respond to the request with a static response. |
binds[].listeners[].routes[].policies.directResponse.body |
|
binds[].listeners[].routes[].policies.directResponse.status |
|
binds[].listeners[].routes[].policies.cors |
Handle CORS preflight requests and append configured CORS headers to applicable requests. |
binds[].listeners[].routes[].policies.cors.allowCredentials |
|
binds[].listeners[].routes[].policies.cors.allowHeaders |
|
binds[].listeners[].routes[].policies.cors.allowMethods |
|
binds[].listeners[].routes[].policies.cors.allowOrigins |
|
binds[].listeners[].routes[].policies.cors.exposeHeaders |
|
binds[].listeners[].routes[].policies.cors.maxAge |
|
binds[].listeners[].routes[].policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].policies.authorization |
Authorization policies for HTTP access. |
binds[].listeners[].routes[].policies.authorization.rules |
|
binds[].listeners[].routes[].policies.mcpAuthentication |
Authentication for MCP clients. |
binds[].listeners[].routes[].policies.mcpAuthentication.issuer |
|
binds[].listeners[].routes[].policies.mcpAuthentication.audiences |
|
binds[].listeners[].routes[].policies.mcpAuthentication.provider |
|
binds[].listeners[].routes[].policies.mcpAuthentication.provider.(any)(1)auth0 |
|
binds[].listeners[].routes[].policies.mcpAuthentication.provider.(any)(1)keycloak |
|
binds[].listeners[].routes[].policies.mcpAuthentication.resourceMetadata |
|
binds[].listeners[].routes[].policies.mcpAuthentication.jwks |
|
binds[].listeners[].routes[].policies.mcpAuthentication.jwks.(any)file |
|
binds[].listeners[].routes[].policies.mcpAuthentication.jwks.(any)url |
|
binds[].listeners[].routes[].policies.mcpAuthentication.mode |
|
binds[].listeners[].routes[].policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].policies.ai.promptGuard |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)regex |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)regex.action |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)regex.rules |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.target |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.status |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.cert |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.root |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.hostname |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.insecure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.alpn |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.http.version |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.http.requestTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.status |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.cert |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.root |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.hostname |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.insecure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.alpn |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http.version |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http.requestTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.status |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.cert |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.root |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.hostname |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.insecure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.alpn |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.http.version |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.http.requestTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].rejection |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].rejection.body |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].rejection.status |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].policies.ai.promptGuard.request[].rejection.headers.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].rejection.headers.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.request[].rejection.headers.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)regex |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)regex.action |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)regex.rules |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.target |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.status |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.cert |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.root |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.hostname |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.insecure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.alpn |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http.version |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http.requestTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.status |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.cert |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.root |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.hostname |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.insecure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.alpn |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.http.version |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.http.requestTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].rejection |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].rejection.body |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].rejection.status |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].policies.ai.promptGuard.response[].rejection.headers.add |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].rejection.headers.set |
|
binds[].listeners[].routes[].policies.ai.promptGuard.response[].rejection.headers.remove |
|
binds[].listeners[].routes[].policies.ai.defaults |
|
binds[].listeners[].routes[].policies.ai.overrides |
|
binds[].listeners[].routes[].policies.ai.transformations |
|
binds[].listeners[].routes[].policies.ai.prompts |
|
binds[].listeners[].routes[].policies.ai.prompts.append |
|
binds[].listeners[].routes[].policies.ai.prompts.append[].role |
|
binds[].listeners[].routes[].policies.ai.prompts.append[].content |
|
binds[].listeners[].routes[].policies.ai.prompts.prepend |
|
binds[].listeners[].routes[].policies.ai.prompts.prepend[].role |
|
binds[].listeners[].routes[].policies.ai.prompts.prepend[].content |
|
binds[].listeners[].routes[].policies.ai.modelAliases |
|
binds[].listeners[].routes[].policies.ai.promptCaching |
|
binds[].listeners[].routes[].policies.ai.promptCaching.cacheSystem |
|
binds[].listeners[].routes[].policies.ai.promptCaching.cacheMessages |
|
binds[].listeners[].routes[].policies.ai.promptCaching.cacheTools |
|
binds[].listeners[].routes[].policies.ai.promptCaching.minTokens |
|
binds[].listeners[].routes[].policies.ai.routes |
|
binds[].listeners[].routes[].policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].policies.backendTLS.cert |
|
binds[].listeners[].routes[].policies.backendTLS.key |
|
binds[].listeners[].routes[].policies.backendTLS.root |
|
binds[].listeners[].routes[].policies.backendTLS.hostname |
|
binds[].listeners[].routes[].policies.backendTLS.insecure |
|
binds[].listeners[].routes[].policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].policies.backendTLS.alpn |
|
binds[].listeners[].routes[].policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].policies.localRateLimit |
Rate limit incoming requests. State is kept local. |
binds[].listeners[].routes[].policies.localRateLimit[].maxTokens |
|
binds[].listeners[].routes[].policies.localRateLimit[].tokensPerFill |
|
binds[].listeners[].routes[].policies.localRateLimit[].fillInterval |
|
binds[].listeners[].routes[].policies.localRateLimit[].type |
|
binds[].listeners[].routes[].policies.remoteRateLimit |
Rate limit incoming requests. State is managed by a remote server. |
binds[].listeners[].routes[].policies.remoteRateLimit.(any)(1)service |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)(1)service.name |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)(1)service.name.namespace |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)(1)service.name.hostname |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)(1)service.port |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)(1)host |
Hostname or IP address |
binds[].listeners[].routes[].policies.remoteRateLimit.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].policies.remoteRateLimit.(any)domain |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)descriptors |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)descriptors[].entries |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)descriptors[].entries[].key |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)descriptors[].entries[].value |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)descriptors[].type |
|
binds[].listeners[].routes[].policies.remoteRateLimit.(any)timeout |
Timeout for the request |
binds[].listeners[].routes[].policies.jwtAuth |
Authenticate incoming JWT requests. |
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)mode |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)providers |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)providers[].issuer |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)providers[].audiences |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)providers[].jwks |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)providers[].jwks.(any)file |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)providers[].jwks.(any)url |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)mode |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)issuer |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)audiences |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)jwks |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)jwks.(any)file |
|
binds[].listeners[].routes[].policies.jwtAuth.(any)(any)jwks.(any)url |
|
binds[].listeners[].routes[].policies.basicAuth |
Authenticate incoming requests using Basic Authentication with htpasswd. |
binds[].listeners[].routes[].policies.basicAuth.htpasswd |
.htpasswd file contents/reference |
binds[].listeners[].routes[].policies.basicAuth.htpasswd.(any)file |
|
binds[].listeners[].routes[].policies.basicAuth.realm |
Realm name for the WWW-Authenticate header |
binds[].listeners[].routes[].policies.basicAuth.mode |
Validation mode for basic authentication |
binds[].listeners[].routes[].policies.apiKey |
Authenticate incoming requests using API Keys |
binds[].listeners[].routes[].policies.apiKey.keys |
List of API keys |
binds[].listeners[].routes[].policies.apiKey.keys[].key |
|
binds[].listeners[].routes[].policies.apiKey.keys[].metadata |
|
binds[].listeners[].routes[].policies.apiKey.mode |
Validation mode for API keys |
binds[].listeners[].routes[].policies.extAuthz |
Authenticate incoming requests by calling an external authorization server. |
binds[].listeners[].routes[].policies.extAuthz.(any)(1)service |
|
binds[].listeners[].routes[].policies.extAuthz.(any)(1)service.name |
|
binds[].listeners[].routes[].policies.extAuthz.(any)(1)service.name.namespace |
|
binds[].listeners[].routes[].policies.extAuthz.(any)(1)service.name.hostname |
|
binds[].listeners[].routes[].policies.extAuthz.(any)(1)service.port |
|
binds[].listeners[].routes[].policies.extAuthz.(any)(1)host |
Hostname or IP address |
binds[].listeners[].routes[].policies.extAuthz.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].policies.extAuthz.(any)protocol |
The ext_authz protocol to use. Unless you need to integrate with an HTTP-only server, gRPC is recommended. |
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)grpc |
|
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)grpc.context |
Additional context to send to the authorization service. This maps to the context_extensions field of the request, and only allows static values. |
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)grpc.metadata |
Additional metadata to send to the authorization service. This maps to the metadata_context.filter_metadata field of the request, and allows dynamic CEL expressions.If unset, by default the envoy.filters.http.jwt_authn key is set if the JWT policy is used as well, for compatibility. |
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)http |
|
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)http.path |
|
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)http.redirect |
When using the HTTP protocol, and the server returns unauthorized, redirect to the URL resolved by the provided expression rather than directly returning the error. |
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)http.includeResponseHeaders |
Specific headers from the authorization response will be copied into the request to the backend. |
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)http.addRequestHeaders |
Specific headers to add in the authorization request (empty = all headers), based on the expression |
binds[].listeners[].routes[].policies.extAuthz.(any)protocol.(1)http.metadata |
Metadata to include under the extauthz variable, based on the authorization response. |
binds[].listeners[].routes[].policies.extAuthz.(any)failureMode |
Behavior when the authorization service is unavailable or returns an error |
binds[].listeners[].routes[].policies.extAuthz.(any)failureMode.(1)denyWithStatus |
|
binds[].listeners[].routes[].policies.extAuthz.(any)includeRequestHeaders |
Specific headers to include in the authorization request. If unset, the gRPC protocol sends all request headers. The HTTP protocol sends only ‘Authorization’. |
binds[].listeners[].routes[].policies.extAuthz.(any)includeRequestBody |
Options for including the request body in the authorization request |
binds[].listeners[].routes[].policies.extAuthz.(any)includeRequestBody.maxRequestBytes |
Maximum size of request body to buffer (default: 8192) |
binds[].listeners[].routes[].policies.extAuthz.(any)includeRequestBody.allowPartialMessage |
If true, send partial body when max_request_bytes is reached |
binds[].listeners[].routes[].policies.extAuthz.(any)includeRequestBody.packAsBytes |
If true, pack body as raw bytes in gRPC |
binds[].listeners[].routes[].policies.extAuthz.(any)timeout |
Timeout for the authorization request (default: 200ms) |
binds[].listeners[].routes[].policies.extProc |
Extend agentgateway with an external processor |
binds[].listeners[].routes[].policies.extProc.(any)(1)service |
|
binds[].listeners[].routes[].policies.extProc.(any)(1)service.name |
|
binds[].listeners[].routes[].policies.extProc.(any)(1)service.name.namespace |
|
binds[].listeners[].routes[].policies.extProc.(any)(1)service.name.hostname |
|
binds[].listeners[].routes[].policies.extProc.(any)(1)service.port |
|
binds[].listeners[].routes[].policies.extProc.(any)(1)host |
Hostname or IP address |
binds[].listeners[].routes[].policies.extProc.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].policies.extProc.(any)failureMode |
Behavior when the ext_proc service is unavailable or returns an error |
binds[].listeners[].routes[].policies.extProc.(any)metadataContext |
Additional metadata to send to the external processing service. Maps to the metadata_context.filter_metadata field in ProcessingRequest, and allows dynamic CEL expressions. |
binds[].listeners[].routes[].policies.extProc.(any)requestAttributes |
Maps to the request attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
binds[].listeners[].routes[].policies.extProc.(any)responseAttributes |
Maps to the response attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
binds[].listeners[].routes[].policies.transformations |
Modify requests and responses |
binds[].listeners[].routes[].policies.transformations.request |
|
binds[].listeners[].routes[].policies.transformations.request.add |
|
binds[].listeners[].routes[].policies.transformations.request.set |
|
binds[].listeners[].routes[].policies.transformations.request.remove |
|
binds[].listeners[].routes[].policies.transformations.request.body |
|
binds[].listeners[].routes[].policies.transformations.response |
|
binds[].listeners[].routes[].policies.transformations.response.add |
|
binds[].listeners[].routes[].policies.transformations.response.set |
|
binds[].listeners[].routes[].policies.transformations.response.remove |
|
binds[].listeners[].routes[].policies.transformations.response.body |
|
binds[].listeners[].routes[].policies.csrf |
Handle CSRF protection by validating request origins against configured allowed origins. |
binds[].listeners[].routes[].policies.csrf.additionalOrigins |
|
binds[].listeners[].routes[].policies.timeout |
Timeout requests that exceed the configured duration. |
binds[].listeners[].routes[].policies.timeout.requestTimeout |
|
binds[].listeners[].routes[].policies.timeout.backendRequestTimeout |
|
binds[].listeners[].routes[].policies.retry |
Retry matching requests. |
binds[].listeners[].routes[].policies.retry.attempts |
|
binds[].listeners[].routes[].policies.retry.backoff |
|
binds[].listeners[].routes[].policies.retry.codes |
|
binds[].listeners[].routes[].backends |
|
binds[].listeners[].routes[].backends[].(1)service |
|
binds[].listeners[].routes[].backends[].(1)service.name |
|
binds[].listeners[].routes[].backends[].(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].(1)service.port |
|
binds[].listeners[].routes[].backends[].(1)host |
|
binds[].listeners[].routes[].backends[].(1)dynamic |
|
binds[].listeners[].routes[].backends[].(1)mcp |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)sse |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)sse.host |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)sse.port |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)sse.path |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)mcp |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)mcp.host |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)mcp.port |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)mcp.path |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)stdio |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)stdio.cmd |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)stdio.args |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)stdio.env |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)openapi |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)openapi.host |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)openapi.port |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)openapi.path |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].(1)openapi.schema |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].name |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.authority |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.path |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.requestRedirect.status |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)regex |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)regex.action |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)regex.rules |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)openAIModeration |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)openAIModeration.policies |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].rejection |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].rejection.body |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].rejection.status |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].rejection.headers.add |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].rejection.headers.set |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.request[].rejection.headers.remove |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)regex |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)regex.action |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)regex.rules |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].rejection |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].rejection.body |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].rejection.status |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].rejection.headers.add |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].rejection.headers.set |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptGuard.response[].rejection.headers.remove |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.defaults |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.overrides |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.transformations |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.prompts |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.prompts.append |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.prompts.append[].role |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.prompts.append[].content |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.prompts.prepend |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.prompts.prepend[].role |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.prompts.prepend[].content |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.modelAliases |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptCaching |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptCaching.cacheSystem |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptCaching.cacheMessages |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptCaching.cacheTools |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.promptCaching.minTokens |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.ai.routes |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS.cert |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS.key |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS.root |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS.hostname |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS.insecure |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS.alpn |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.http.version |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.http.requestTimeout |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp.keepalives |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].backends[].(1)mcp.targets[].policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].routes[].backends[].(1)mcp.statefulMode |
|
binds[].listeners[].routes[].backends[].(1)mcp.prefixMode |
|
binds[].listeners[].routes[].backends[].(1)ai |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)openAI |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)openAI.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)gemini |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)gemini.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)vertex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)vertex.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)vertex.region |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)vertex.projectId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)anthropic |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)anthropic.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)bedrock |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)bedrock.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)bedrock.region |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)bedrock.guardrailIdentifier |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)bedrock.guardrailVersion |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)azureOpenAI |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)azureOpenAI.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)azureOpenAI.host |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)provider.(1)azureOpenAI.apiVersion |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)hostOverride |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)pathOverride |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)tokenize |
Whether to tokenize on the request flow. This enables us to do more accurate rate limits, since we know (part of) the cost of the request upfront. This comes with the cost of an expensive operation. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.authority |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.path |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.requestRedirect.status |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)regex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)regex.action |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)regex.rules |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.target |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)openAIModeration |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)openAIModeration.policies |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].rejection |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].rejection.body |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].rejection.status |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].rejection.headers.add |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].rejection.headers.set |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.request[].rejection.headers.remove |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)regex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)regex.action |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)regex.rules |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.target |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].rejection |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].rejection.body |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].rejection.status |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].rejection.headers.add |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].rejection.headers.set |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptGuard.response[].rejection.headers.remove |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.defaults |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.overrides |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.transformations |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.prompts |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.prompts.append |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.prompts.append[].role |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.prompts.append[].content |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.prompts.prepend |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.prompts.prepend[].role |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.prompts.prepend[].content |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.modelAliases |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptCaching |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptCaching.cacheSystem |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptCaching.cacheMessages |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptCaching.cacheTools |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.promptCaching.minTokens |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.ai.routes |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS.cert |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS.key |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS.root |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS.hostname |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS.insecure |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS.alpn |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.http.version |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.http.requestTimeout |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp.keepalives |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)openAI |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)openAI.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)gemini |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)gemini.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)vertex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)vertex.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)vertex.region |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)vertex.projectId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)anthropic |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)anthropic.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)bedrock |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)bedrock.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)bedrock.region |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)bedrock.guardrailIdentifier |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)bedrock.guardrailVersion |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)azureOpenAI |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)azureOpenAI.model |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)azureOpenAI.host |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].provider.(1)azureOpenAI.apiVersion |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].hostOverride |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].pathOverride |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].tokenize |
Whether to tokenize on the request flow. This enables us to do more accurate rate limits, since we know (part of) the cost of the request upfront. This comes with the cost of an expensive operation. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.authority |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.path |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.requestRedirect.status |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)regex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)regex.action |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)regex.rules |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.target |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)openAIModeration |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)openAIModeration.policies |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].rejection |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].rejection.body |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].rejection.status |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].rejection.headers.add |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].rejection.headers.set |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.request[].rejection.headers.remove |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)regex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)regex.action |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)regex.rules |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.target |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].rejection |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].rejection.body |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].rejection.status |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].rejection.headers.add |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].rejection.headers.set |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptGuard.response[].rejection.headers.remove |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.defaults |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.overrides |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.transformations |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.prompts |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.prompts.append |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.prompts.append[].role |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.prompts.append[].content |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.prompts.prepend |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.prompts.prepend[].role |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.prompts.prepend[].content |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.modelAliases |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptCaching |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptCaching.cacheSystem |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptCaching.cacheMessages |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptCaching.cacheTools |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.promptCaching.minTokens |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.ai.routes |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS.cert |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS.key |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS.root |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS.hostname |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS.insecure |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS.alpn |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.http.version |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.http.requestTimeout |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp.keepalives |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].backends[].(1)ai.(any)groups[].providers[].policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].routes[].backends[].weight |
|
binds[].listeners[].routes[].backends[].policies |
|
binds[].listeners[].routes[].backends[].policies.requestHeaderModifier |
Headers to be modified in the request. |
binds[].listeners[].routes[].backends[].policies.requestHeaderModifier.add |
|
binds[].listeners[].routes[].backends[].policies.requestHeaderModifier.set |
|
binds[].listeners[].routes[].backends[].policies.requestHeaderModifier.remove |
|
binds[].listeners[].routes[].backends[].policies.responseHeaderModifier |
Headers to be modified in the response. |
binds[].listeners[].routes[].backends[].policies.responseHeaderModifier.add |
|
binds[].listeners[].routes[].backends[].policies.responseHeaderModifier.set |
|
binds[].listeners[].routes[].backends[].policies.responseHeaderModifier.remove |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect |
Directly respond to the request with a redirect. |
binds[].listeners[].routes[].backends[].policies.requestRedirect.scheme |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect.authority |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect.authority.(any)(1)full |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect.authority.(any)(1)host |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect.authority.(any)(1)port |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect.path |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect.path.(any)(1)full |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect.path.(any)(1)prefix |
|
binds[].listeners[].routes[].backends[].policies.requestRedirect.status |
|
binds[].listeners[].routes[].backends[].policies.mcpAuthorization |
Authorization policies for MCP access. |
binds[].listeners[].routes[].backends[].policies.mcpAuthorization.rules |
|
binds[].listeners[].routes[].backends[].policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
binds[].listeners[].routes[].backends[].policies.ai |
Mark this as LLM traffic to enable LLM processing. |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)regex |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)regex.action |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)regex.rules |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.target |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)openAIModeration |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)openAIModeration.policies |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].rejection |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].rejection.body |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].rejection.status |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].rejection.headers.add |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].rejection.headers.set |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.request[].rejection.headers.remove |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)regex |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)regex.action |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)regex.rules |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.target |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].rejection |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].rejection.body |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].rejection.status |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].rejection.headers.add |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].rejection.headers.set |
|
binds[].listeners[].routes[].backends[].policies.ai.promptGuard.response[].rejection.headers.remove |
|
binds[].listeners[].routes[].backends[].policies.ai.defaults |
|
binds[].listeners[].routes[].backends[].policies.ai.overrides |
|
binds[].listeners[].routes[].backends[].policies.ai.transformations |
|
binds[].listeners[].routes[].backends[].policies.ai.prompts |
|
binds[].listeners[].routes[].backends[].policies.ai.prompts.append |
|
binds[].listeners[].routes[].backends[].policies.ai.prompts.append[].role |
|
binds[].listeners[].routes[].backends[].policies.ai.prompts.append[].content |
|
binds[].listeners[].routes[].backends[].policies.ai.prompts.prepend |
|
binds[].listeners[].routes[].backends[].policies.ai.prompts.prepend[].role |
|
binds[].listeners[].routes[].backends[].policies.ai.prompts.prepend[].content |
|
binds[].listeners[].routes[].backends[].policies.ai.modelAliases |
|
binds[].listeners[].routes[].backends[].policies.ai.promptCaching |
|
binds[].listeners[].routes[].backends[].policies.ai.promptCaching.cacheSystem |
|
binds[].listeners[].routes[].backends[].policies.ai.promptCaching.cacheMessages |
|
binds[].listeners[].routes[].backends[].policies.ai.promptCaching.cacheTools |
|
binds[].listeners[].routes[].backends[].policies.ai.promptCaching.minTokens |
|
binds[].listeners[].routes[].backends[].policies.ai.routes |
|
binds[].listeners[].routes[].backends[].policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].routes[].backends[].policies.backendTLS.cert |
|
binds[].listeners[].routes[].backends[].policies.backendTLS.key |
|
binds[].listeners[].routes[].backends[].policies.backendTLS.root |
|
binds[].listeners[].routes[].backends[].policies.backendTLS.hostname |
|
binds[].listeners[].routes[].backends[].policies.backendTLS.insecure |
|
binds[].listeners[].routes[].backends[].policies.backendTLS.insecureHost |
|
binds[].listeners[].routes[].backends[].policies.backendTLS.alpn |
|
binds[].listeners[].routes[].backends[].policies.backendTLS.subjectAltNames |
|
binds[].listeners[].routes[].backends[].policies.backendAuth |
Authenticate to the backend. |
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)passthrough |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)key |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)key.(any)file |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)gcp |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)gcp.(any)type |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)aws |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)aws.(any)region |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
binds[].listeners[].routes[].backends[].policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
binds[].listeners[].routes[].backends[].policies.http |
Specify HTTP settings for the backend |
binds[].listeners[].routes[].backends[].policies.http.version |
|
binds[].listeners[].routes[].backends[].policies.http.requestTimeout |
|
binds[].listeners[].routes[].backends[].policies.tcp |
Specify TCP settings for the backend |
binds[].listeners[].routes[].backends[].policies.tcp.keepalives |
|
binds[].listeners[].routes[].backends[].policies.tcp.keepalives.enabled |
|
binds[].listeners[].routes[].backends[].policies.tcp.keepalives.time |
|
binds[].listeners[].routes[].backends[].policies.tcp.keepalives.interval |
|
binds[].listeners[].routes[].backends[].policies.tcp.keepalives.retries |
|
binds[].listeners[].routes[].backends[].policies.tcp.connectTimeout |
|
binds[].listeners[].routes[].backends[].policies.tcp.connectTimeout.secs |
|
binds[].listeners[].routes[].backends[].policies.tcp.connectTimeout.nanos |
|
binds[].listeners[].tcpRoutes |
|
binds[].listeners[].tcpRoutes[].name |
|
binds[].listeners[].tcpRoutes[].namespace |
|
binds[].listeners[].tcpRoutes[].ruleName |
|
binds[].listeners[].tcpRoutes[].hostnames |
Can be a wildcard |
binds[].listeners[].tcpRoutes[].policies |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS.cert |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS.key |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS.root |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS.hostname |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS.insecure |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS.insecureHost |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS.alpn |
|
binds[].listeners[].tcpRoutes[].policies.backendTLS.subjectAltNames |
|
binds[].listeners[].tcpRoutes[].backends |
|
binds[].listeners[].tcpRoutes[].backends[].(1)service |
|
binds[].listeners[].tcpRoutes[].backends[].(1)service.name |
|
binds[].listeners[].tcpRoutes[].backends[].(1)service.name.namespace |
|
binds[].listeners[].tcpRoutes[].backends[].(1)service.name.hostname |
|
binds[].listeners[].tcpRoutes[].backends[].(1)service.port |
|
binds[].listeners[].tcpRoutes[].backends[].(1)host |
Hostname or IP address |
binds[].listeners[].tcpRoutes[].backends[].(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].tcpRoutes[].backends[].weight |
|
binds[].listeners[].tcpRoutes[].backends[].policies |
|
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS |
Send TLS to the backend. |
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS.cert |
|
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS.key |
|
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS.root |
|
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS.hostname |
|
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS.insecure |
|
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS.insecureHost |
|
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS.alpn |
|
binds[].listeners[].tcpRoutes[].backends[].policies.backendTLS.subjectAltNames |
|
binds[].listeners[].policies |
|
binds[].listeners[].policies.jwtAuth |
Authenticate incoming JWT requests. |
binds[].listeners[].policies.jwtAuth.(any)(any)mode |
|
binds[].listeners[].policies.jwtAuth.(any)(any)providers |
|
binds[].listeners[].policies.jwtAuth.(any)(any)providers[].issuer |
|
binds[].listeners[].policies.jwtAuth.(any)(any)providers[].audiences |
|
binds[].listeners[].policies.jwtAuth.(any)(any)providers[].jwks |
|
binds[].listeners[].policies.jwtAuth.(any)(any)providers[].jwks.(any)file |
|
binds[].listeners[].policies.jwtAuth.(any)(any)providers[].jwks.(any)url |
|
binds[].listeners[].policies.jwtAuth.(any)(any)mode |
|
binds[].listeners[].policies.jwtAuth.(any)(any)issuer |
|
binds[].listeners[].policies.jwtAuth.(any)(any)audiences |
|
binds[].listeners[].policies.jwtAuth.(any)(any)jwks |
|
binds[].listeners[].policies.jwtAuth.(any)(any)jwks.(any)file |
|
binds[].listeners[].policies.jwtAuth.(any)(any)jwks.(any)url |
|
binds[].listeners[].policies.extAuthz |
Authenticate incoming requests by calling an external authorization server. |
binds[].listeners[].policies.extAuthz.(any)(1)service |
|
binds[].listeners[].policies.extAuthz.(any)(1)service.name |
|
binds[].listeners[].policies.extAuthz.(any)(1)service.name.namespace |
|
binds[].listeners[].policies.extAuthz.(any)(1)service.name.hostname |
|
binds[].listeners[].policies.extAuthz.(any)(1)service.port |
|
binds[].listeners[].policies.extAuthz.(any)(1)host |
Hostname or IP address |
binds[].listeners[].policies.extAuthz.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].policies.extAuthz.(any)protocol |
The ext_authz protocol to use. Unless you need to integrate with an HTTP-only server, gRPC is recommended. |
binds[].listeners[].policies.extAuthz.(any)protocol.(1)grpc |
|
binds[].listeners[].policies.extAuthz.(any)protocol.(1)grpc.context |
Additional context to send to the authorization service. This maps to the context_extensions field of the request, and only allows static values. |
binds[].listeners[].policies.extAuthz.(any)protocol.(1)grpc.metadata |
Additional metadata to send to the authorization service. This maps to the metadata_context.filter_metadata field of the request, and allows dynamic CEL expressions.If unset, by default the envoy.filters.http.jwt_authn key is set if the JWT policy is used as well, for compatibility. |
binds[].listeners[].policies.extAuthz.(any)protocol.(1)http |
|
binds[].listeners[].policies.extAuthz.(any)protocol.(1)http.path |
|
binds[].listeners[].policies.extAuthz.(any)protocol.(1)http.redirect |
When using the HTTP protocol, and the server returns unauthorized, redirect to the URL resolved by the provided expression rather than directly returning the error. |
binds[].listeners[].policies.extAuthz.(any)protocol.(1)http.includeResponseHeaders |
Specific headers from the authorization response will be copied into the request to the backend. |
binds[].listeners[].policies.extAuthz.(any)protocol.(1)http.addRequestHeaders |
Specific headers to add in the authorization request (empty = all headers), based on the expression |
binds[].listeners[].policies.extAuthz.(any)protocol.(1)http.metadata |
Metadata to include under the extauthz variable, based on the authorization response. |
binds[].listeners[].policies.extAuthz.(any)failureMode |
Behavior when the authorization service is unavailable or returns an error |
binds[].listeners[].policies.extAuthz.(any)failureMode.(1)denyWithStatus |
|
binds[].listeners[].policies.extAuthz.(any)includeRequestHeaders |
Specific headers to include in the authorization request. If unset, the gRPC protocol sends all request headers. The HTTP protocol sends only ‘Authorization’. |
binds[].listeners[].policies.extAuthz.(any)includeRequestBody |
Options for including the request body in the authorization request |
binds[].listeners[].policies.extAuthz.(any)includeRequestBody.maxRequestBytes |
Maximum size of request body to buffer (default: 8192) |
binds[].listeners[].policies.extAuthz.(any)includeRequestBody.allowPartialMessage |
If true, send partial body when max_request_bytes is reached |
binds[].listeners[].policies.extAuthz.(any)includeRequestBody.packAsBytes |
If true, pack body as raw bytes in gRPC |
binds[].listeners[].policies.extAuthz.(any)timeout |
Timeout for the authorization request (default: 200ms) |
binds[].listeners[].policies.extProc |
Extend agentgateway with an external processor |
binds[].listeners[].policies.extProc.(any)(1)service |
|
binds[].listeners[].policies.extProc.(any)(1)service.name |
|
binds[].listeners[].policies.extProc.(any)(1)service.name.namespace |
|
binds[].listeners[].policies.extProc.(any)(1)service.name.hostname |
|
binds[].listeners[].policies.extProc.(any)(1)service.port |
|
binds[].listeners[].policies.extProc.(any)(1)host |
Hostname or IP address |
binds[].listeners[].policies.extProc.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
binds[].listeners[].policies.extProc.(any)failureMode |
Behavior when the ext_proc service is unavailable or returns an error |
binds[].listeners[].policies.extProc.(any)metadataContext |
Additional metadata to send to the external processing service. Maps to the metadata_context.filter_metadata field in ProcessingRequest, and allows dynamic CEL expressions. |
binds[].listeners[].policies.extProc.(any)requestAttributes |
Maps to the request attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
binds[].listeners[].policies.extProc.(any)responseAttributes |
Maps to the response attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
binds[].listeners[].policies.transformations |
Modify requests and responses |
binds[].listeners[].policies.transformations.request |
|
binds[].listeners[].policies.transformations.request.add |
|
binds[].listeners[].policies.transformations.request.set |
|
binds[].listeners[].policies.transformations.request.remove |
|
binds[].listeners[].policies.transformations.request.body |
|
binds[].listeners[].policies.transformations.response |
|
binds[].listeners[].policies.transformations.response.add |
|
binds[].listeners[].policies.transformations.response.set |
|
binds[].listeners[].policies.transformations.response.remove |
|
binds[].listeners[].policies.transformations.response.body |
|
binds[].listeners[].policies.basicAuth |
Authenticate incoming requests using Basic Authentication with htpasswd. |
binds[].listeners[].policies.basicAuth.htpasswd |
.htpasswd file contents/reference |
binds[].listeners[].policies.basicAuth.htpasswd.(any)file |
|
binds[].listeners[].policies.basicAuth.realm |
Realm name for the WWW-Authenticate header |
binds[].listeners[].policies.basicAuth.mode |
Validation mode for basic authentication |
binds[].listeners[].policies.apiKey |
Authenticate incoming requests using API Keys |
binds[].listeners[].policies.apiKey.keys |
List of API keys |
binds[].listeners[].policies.apiKey.keys[].key |
|
binds[].listeners[].policies.apiKey.keys[].metadata |
|
binds[].listeners[].policies.apiKey.mode |
Validation mode for API keys |
binds[].tunnelProtocol |
|
frontendPolicies |
|
frontendPolicies.http |
Settings for handling incoming HTTP requests. |
frontendPolicies.http.maxBufferSize |
|
frontendPolicies.http.http1MaxHeaders |
The maximum number of headers allowed in a request. Changing this value results in a performance degradation, even if set to a lower value than the default (100) |
frontendPolicies.http.http1IdleTimeout |
|
frontendPolicies.http.http2WindowSize |
|
frontendPolicies.http.http2ConnectionWindowSize |
|
frontendPolicies.http.http2FrameSize |
|
frontendPolicies.http.http2KeepaliveInterval |
|
frontendPolicies.http.http2KeepaliveTimeout |
|
frontendPolicies.tls |
Settings for handling incoming TLS connections. |
frontendPolicies.tls.handshakeTimeout |
|
frontendPolicies.tls.alpn |
|
frontendPolicies.tls.minVersion |
|
frontendPolicies.tls.maxVersion |
|
frontendPolicies.tls.cipherSuites |
|
frontendPolicies.tcp |
Settings for handling incoming TCP connections. |
frontendPolicies.tcp.keepalives |
|
frontendPolicies.tcp.keepalives.enabled |
|
frontendPolicies.tcp.keepalives.time |
|
frontendPolicies.tcp.keepalives.interval |
|
frontendPolicies.tcp.keepalives.retries |
|
frontendPolicies.accessLog |
Settings for request access logs. |
frontendPolicies.accessLog.filter |
|
frontendPolicies.accessLog.add |
|
frontendPolicies.accessLog.remove |
|
frontendPolicies.tracing |
|
frontendPolicies.tracing.(any)(1)service |
|
frontendPolicies.tracing.(any)(1)service.name |
|
frontendPolicies.tracing.(any)(1)service.name.namespace |
|
frontendPolicies.tracing.(any)(1)service.name.hostname |
|
frontendPolicies.tracing.(any)(1)service.port |
|
frontendPolicies.tracing.(any)(1)host |
Hostname or IP address |
frontendPolicies.tracing.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
frontendPolicies.tracing.(any)attributes |
Span attributes to add, keyed by attribute name. |
frontendPolicies.tracing.(any)resources |
Resource attributes to add to the tracer provider (OTel Resource).This can be used to set things like service.name dynamically. |
frontendPolicies.tracing.(any)remove |
Attribute keys to remove from the emitted span attributes. This is applied before attributes are evaluated/added, so it can be used to dropdefault attributes or avoid duplication. |
frontendPolicies.tracing.(any)randomSampling |
Optional per-policy override for random sampling. If set, overrides global config for requests that use this frontend policy. |
frontendPolicies.tracing.(any)clientSampling |
Optional per-policy override for client sampling. If set, overrides global config for requests that use this frontend policy. |
frontendPolicies.tracing.(any)path |
|
frontendPolicies.tracing.(any)protocol |
|
policies |
policies defines additional policies that can be attached to various other configurations. This is an advanced feature; users should typically use the inline policies field under route/gateway. |
policies[].name |
|
policies[].name.name |
|
policies[].name.namespace |
|
policies[].target |
|
policies[].target.(1)gateway |
|
policies[].target.(1)gateway.gatewayName |
|
policies[].target.(1)gateway.gatewayNamespace |
|
policies[].target.(1)gateway.listenerName |
|
policies[].target.(1)route |
|
policies[].target.(1)route.name |
|
policies[].target.(1)route.namespace |
|
policies[].target.(1)route.ruleName |
|
policies[].target.(1)route.kind |
|
policies[].target.(1)backend |
|
policies[].target.(1)backend.(1)backend |
|
policies[].target.(1)backend.(1)backend.name |
|
policies[].target.(1)backend.(1)backend.namespace |
|
policies[].target.(1)backend.(1)backend.section |
|
policies[].target.(1)backend.(1)service |
|
policies[].target.(1)backend.(1)service.hostname |
|
policies[].target.(1)backend.(1)service.namespace |
|
policies[].target.(1)backend.(1)service.port |
|
policies[].phase |
phase defines at what level the policy runs at. Gateway policies run pre-routing, while Route policies apply post-routing. Only a subset of policies are eligible as Gateway policies. In general, normal (route level) policies should be used, except you need the policy to influence routing. |
policies[].policy |
|
policies[].policy.requestHeaderModifier |
Headers to be modified in the request. |
policies[].policy.requestHeaderModifier.add |
|
policies[].policy.requestHeaderModifier.set |
|
policies[].policy.requestHeaderModifier.remove |
|
policies[].policy.responseHeaderModifier |
Headers to be modified in the response. |
policies[].policy.responseHeaderModifier.add |
|
policies[].policy.responseHeaderModifier.set |
|
policies[].policy.responseHeaderModifier.remove |
|
policies[].policy.requestRedirect |
Directly respond to the request with a redirect. |
policies[].policy.requestRedirect.scheme |
|
policies[].policy.requestRedirect.authority |
|
policies[].policy.requestRedirect.authority.(any)(1)full |
|
policies[].policy.requestRedirect.authority.(any)(1)host |
|
policies[].policy.requestRedirect.authority.(any)(1)port |
|
policies[].policy.requestRedirect.path |
|
policies[].policy.requestRedirect.path.(any)(1)full |
|
policies[].policy.requestRedirect.path.(any)(1)prefix |
|
policies[].policy.requestRedirect.status |
|
policies[].policy.urlRewrite |
Modify the URL path or authority. |
policies[].policy.urlRewrite.authority |
|
policies[].policy.urlRewrite.authority.(any)(1)full |
|
policies[].policy.urlRewrite.authority.(any)(1)host |
|
policies[].policy.urlRewrite.authority.(any)(1)port |
|
policies[].policy.urlRewrite.path |
|
policies[].policy.urlRewrite.path.(any)(1)full |
|
policies[].policy.urlRewrite.path.(any)(1)prefix |
|
policies[].policy.requestMirror |
Mirror incoming requests to another destination. |
policies[].policy.requestMirror.backend |
|
policies[].policy.requestMirror.backend.(1)service |
|
policies[].policy.requestMirror.backend.(1)service.name |
|
policies[].policy.requestMirror.backend.(1)service.name.namespace |
|
policies[].policy.requestMirror.backend.(1)service.name.hostname |
|
policies[].policy.requestMirror.backend.(1)service.port |
|
policies[].policy.requestMirror.backend.(1)host |
Hostname or IP address |
policies[].policy.requestMirror.backend.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
policies[].policy.requestMirror.percentage |
|
policies[].policy.directResponse |
Directly respond to the request with a static response. |
policies[].policy.directResponse.body |
|
policies[].policy.directResponse.status |
|
policies[].policy.cors |
Handle CORS preflight requests and append configured CORS headers to applicable requests. |
policies[].policy.cors.allowCredentials |
|
policies[].policy.cors.allowHeaders |
|
policies[].policy.cors.allowMethods |
|
policies[].policy.cors.allowOrigins |
|
policies[].policy.cors.exposeHeaders |
|
policies[].policy.cors.maxAge |
|
policies[].policy.mcpAuthorization |
Authorization policies for MCP access. |
policies[].policy.mcpAuthorization.rules |
|
policies[].policy.authorization |
Authorization policies for HTTP access. |
policies[].policy.authorization.rules |
|
policies[].policy.mcpAuthentication |
Authentication for MCP clients. |
policies[].policy.mcpAuthentication.issuer |
|
policies[].policy.mcpAuthentication.audiences |
|
policies[].policy.mcpAuthentication.provider |
|
policies[].policy.mcpAuthentication.provider.(any)(1)auth0 |
|
policies[].policy.mcpAuthentication.provider.(any)(1)keycloak |
|
policies[].policy.mcpAuthentication.resourceMetadata |
|
policies[].policy.mcpAuthentication.jwks |
|
policies[].policy.mcpAuthentication.jwks.(any)file |
|
policies[].policy.mcpAuthentication.jwks.(any)url |
|
policies[].policy.mcpAuthentication.mode |
|
policies[].policy.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
policies[].policy.ai |
Mark this as LLM traffic to enable LLM processing. |
policies[].policy.ai.promptGuard |
|
policies[].policy.ai.promptGuard.request |
|
policies[].policy.ai.promptGuard.request[].(1)regex |
|
policies[].policy.ai.promptGuard.request[].(1)regex.action |
|
policies[].policy.ai.promptGuard.request[].(1)regex.rules |
|
policies[].policy.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
policies[].policy.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
policies[].policy.ai.promptGuard.request[].(1)webhook |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.target |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
policies[].policy.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
policies[].policy.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
policies[].policy.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier |
Headers to be modified in the request. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.add |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.set |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier |
Headers to be modified in the response. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.add |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.set |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect |
Directly respond to the request with a redirect. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.scheme |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)full |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)host |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)port |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path.(any)(1)full |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path.(any)(1)prefix |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.status |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.mcpAuthorization |
Authorization policies for MCP access. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.mcpAuthorization.rules |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS |
Send TLS to the backend. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.cert |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.key |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.root |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.hostname |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.insecure |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.insecureHost |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.alpn |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.subjectAltNames |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth |
Authenticate to the backend. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)passthrough |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)key |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)key.(any)file |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)region |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.http |
Specify HTTP settings for the backend |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.http.version |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.http.requestTimeout |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp |
Specify TCP settings for the backend |
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.enabled |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.time |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.interval |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.retries |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout.secs |
|
policies[].policy.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout.nanos |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier |
Headers to be modified in the request. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.add |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.set |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier |
Headers to be modified in the response. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.add |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.set |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect |
Directly respond to the request with a redirect. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.scheme |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)full |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)host |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)port |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)full |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)prefix |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.status |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.mcpAuthorization |
Authorization policies for MCP access. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.mcpAuthorization.rules |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS |
Send TLS to the backend. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.cert |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.key |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.root |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.hostname |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.insecure |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.insecureHost |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.alpn |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.subjectAltNames |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth |
Authenticate to the backend. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)passthrough |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key.(any)file |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)region |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http |
Specify HTTP settings for the backend |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http.version |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http.requestTimeout |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp |
Specify TCP settings for the backend |
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.enabled |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.time |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.interval |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.retries |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout.secs |
|
policies[].policy.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout.nanos |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier |
Headers to be modified in the request. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.add |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.set |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier |
Headers to be modified in the response. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.add |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.set |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect |
Directly respond to the request with a redirect. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.scheme |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)full |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)host |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)port |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)full |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)prefix |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.status |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.mcpAuthorization |
Authorization policies for MCP access. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.mcpAuthorization.rules |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS |
Send TLS to the backend. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.cert |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.key |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.root |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.hostname |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.insecure |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.insecureHost |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.alpn |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.subjectAltNames |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth |
Authenticate to the backend. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)passthrough |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)key |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)key.(any)file |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)region |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.http |
Specify HTTP settings for the backend |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.http.version |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.http.requestTimeout |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp |
Specify TCP settings for the backend |
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.enabled |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.time |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.interval |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.retries |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout.secs |
|
policies[].policy.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout.nanos |
|
policies[].policy.ai.promptGuard.request[].rejection |
|
policies[].policy.ai.promptGuard.request[].rejection.body |
|
policies[].policy.ai.promptGuard.request[].rejection.status |
|
policies[].policy.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
policies[].policy.ai.promptGuard.request[].rejection.headers.add |
|
policies[].policy.ai.promptGuard.request[].rejection.headers.set |
|
policies[].policy.ai.promptGuard.request[].rejection.headers.remove |
|
policies[].policy.ai.promptGuard.response |
|
policies[].policy.ai.promptGuard.response[].(1)regex |
|
policies[].policy.ai.promptGuard.response[].(1)regex.action |
|
policies[].policy.ai.promptGuard.response[].(1)regex.rules |
|
policies[].policy.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
policies[].policy.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
policies[].policy.ai.promptGuard.response[].(1)webhook |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.target |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
policies[].policy.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
policies[].policy.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
policies[].policy.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier |
Headers to be modified in the request. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.add |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.set |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier |
Headers to be modified in the response. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.add |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.set |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect |
Directly respond to the request with a redirect. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.scheme |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)full |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)host |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)port |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)full |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)prefix |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.status |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.mcpAuthorization |
Authorization policies for MCP access. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.mcpAuthorization.rules |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS |
Send TLS to the backend. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.cert |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.key |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.root |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.hostname |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.insecure |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.insecureHost |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.alpn |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.subjectAltNames |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth |
Authenticate to the backend. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)passthrough |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key.(any)file |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)region |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http |
Specify HTTP settings for the backend |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http.version |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http.requestTimeout |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp |
Specify TCP settings for the backend |
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.enabled |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.time |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.interval |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.retries |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout.secs |
|
policies[].policy.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout.nanos |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier |
Headers to be modified in the request. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.add |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.set |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier |
Headers to be modified in the response. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.add |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.set |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.remove |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect |
Directly respond to the request with a redirect. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.scheme |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)full |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)host |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)port |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)full |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)prefix |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.status |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.mcpAuthorization |
Authorization policies for MCP access. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.mcpAuthorization.rules |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS |
Send TLS to the backend. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.cert |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.key |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.root |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.hostname |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.insecure |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.insecureHost |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.alpn |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.subjectAltNames |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth |
Authenticate to the backend. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)passthrough |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)key |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)key.(any)file |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)region |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.http |
Specify HTTP settings for the backend |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.http.version |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.http.requestTimeout |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp |
Specify TCP settings for the backend |
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.enabled |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.time |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.interval |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.retries |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout.secs |
|
policies[].policy.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout.nanos |
|
policies[].policy.ai.promptGuard.response[].rejection |
|
policies[].policy.ai.promptGuard.response[].rejection.body |
|
policies[].policy.ai.promptGuard.response[].rejection.status |
|
policies[].policy.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
policies[].policy.ai.promptGuard.response[].rejection.headers.add |
|
policies[].policy.ai.promptGuard.response[].rejection.headers.set |
|
policies[].policy.ai.promptGuard.response[].rejection.headers.remove |
|
policies[].policy.ai.defaults |
|
policies[].policy.ai.overrides |
|
policies[].policy.ai.transformations |
|
policies[].policy.ai.prompts |
|
policies[].policy.ai.prompts.append |
|
policies[].policy.ai.prompts.append[].role |
|
policies[].policy.ai.prompts.append[].content |
|
policies[].policy.ai.prompts.prepend |
|
policies[].policy.ai.prompts.prepend[].role |
|
policies[].policy.ai.prompts.prepend[].content |
|
policies[].policy.ai.modelAliases |
|
policies[].policy.ai.promptCaching |
|
policies[].policy.ai.promptCaching.cacheSystem |
|
policies[].policy.ai.promptCaching.cacheMessages |
|
policies[].policy.ai.promptCaching.cacheTools |
|
policies[].policy.ai.promptCaching.minTokens |
|
policies[].policy.ai.routes |
|
policies[].policy.backendTLS |
Send TLS to the backend. |
policies[].policy.backendTLS.cert |
|
policies[].policy.backendTLS.key |
|
policies[].policy.backendTLS.root |
|
policies[].policy.backendTLS.hostname |
|
policies[].policy.backendTLS.insecure |
|
policies[].policy.backendTLS.insecureHost |
|
policies[].policy.backendTLS.alpn |
|
policies[].policy.backendTLS.subjectAltNames |
|
policies[].policy.backendAuth |
Authenticate to the backend. |
policies[].policy.backendAuth.(any)(1)passthrough |
|
policies[].policy.backendAuth.(any)(1)key |
|
policies[].policy.backendAuth.(any)(1)key.(any)file |
|
policies[].policy.backendAuth.(any)(1)gcp |
|
policies[].policy.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
policies[].policy.backendAuth.(any)(1)gcp.(any)type |
|
policies[].policy.backendAuth.(any)(1)aws |
|
policies[].policy.backendAuth.(any)(1)aws.(any)accessKeyId |
|
policies[].policy.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
policies[].policy.backendAuth.(any)(1)aws.(any)region |
|
policies[].policy.backendAuth.(any)(1)aws.(any)sessionToken |
|
policies[].policy.backendAuth.(any)(1)azure |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
policies[].policy.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
policies[].policy.backendAuth.(any)(1)azure.(1)developerImplicit |
|
policies[].policy.localRateLimit |
Rate limit incoming requests. State is kept local. |
policies[].policy.localRateLimit[].maxTokens |
|
policies[].policy.localRateLimit[].tokensPerFill |
|
policies[].policy.localRateLimit[].fillInterval |
|
policies[].policy.localRateLimit[].type |
|
policies[].policy.remoteRateLimit |
Rate limit incoming requests. State is managed by a remote server. |
policies[].policy.remoteRateLimit.(any)(1)service |
|
policies[].policy.remoteRateLimit.(any)(1)service.name |
|
policies[].policy.remoteRateLimit.(any)(1)service.name.namespace |
|
policies[].policy.remoteRateLimit.(any)(1)service.name.hostname |
|
policies[].policy.remoteRateLimit.(any)(1)service.port |
|
policies[].policy.remoteRateLimit.(any)(1)host |
Hostname or IP address |
policies[].policy.remoteRateLimit.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
policies[].policy.remoteRateLimit.(any)domain |
|
policies[].policy.remoteRateLimit.(any)descriptors |
|
policies[].policy.remoteRateLimit.(any)descriptors[].entries |
|
policies[].policy.remoteRateLimit.(any)descriptors[].entries[].key |
|
policies[].policy.remoteRateLimit.(any)descriptors[].entries[].value |
|
policies[].policy.remoteRateLimit.(any)descriptors[].type |
|
policies[].policy.remoteRateLimit.(any)timeout |
Timeout for the request |
policies[].policy.jwtAuth |
Authenticate incoming JWT requests. |
policies[].policy.jwtAuth.(any)(any)mode |
|
policies[].policy.jwtAuth.(any)(any)providers |
|
policies[].policy.jwtAuth.(any)(any)providers[].issuer |
|
policies[].policy.jwtAuth.(any)(any)providers[].audiences |
|
policies[].policy.jwtAuth.(any)(any)providers[].jwks |
|
policies[].policy.jwtAuth.(any)(any)providers[].jwks.(any)file |
|
policies[].policy.jwtAuth.(any)(any)providers[].jwks.(any)url |
|
policies[].policy.jwtAuth.(any)(any)mode |
|
policies[].policy.jwtAuth.(any)(any)issuer |
|
policies[].policy.jwtAuth.(any)(any)audiences |
|
policies[].policy.jwtAuth.(any)(any)jwks |
|
policies[].policy.jwtAuth.(any)(any)jwks.(any)file |
|
policies[].policy.jwtAuth.(any)(any)jwks.(any)url |
|
policies[].policy.basicAuth |
Authenticate incoming requests using Basic Authentication with htpasswd. |
policies[].policy.basicAuth.htpasswd |
.htpasswd file contents/reference |
policies[].policy.basicAuth.htpasswd.(any)file |
|
policies[].policy.basicAuth.realm |
Realm name for the WWW-Authenticate header |
policies[].policy.basicAuth.mode |
Validation mode for basic authentication |
policies[].policy.apiKey |
Authenticate incoming requests using API Keys |
policies[].policy.apiKey.keys |
List of API keys |
policies[].policy.apiKey.keys[].key |
|
policies[].policy.apiKey.keys[].metadata |
|
policies[].policy.apiKey.mode |
Validation mode for API keys |
policies[].policy.extAuthz |
Authenticate incoming requests by calling an external authorization server. |
policies[].policy.extAuthz.(any)(1)service |
|
policies[].policy.extAuthz.(any)(1)service.name |
|
policies[].policy.extAuthz.(any)(1)service.name.namespace |
|
policies[].policy.extAuthz.(any)(1)service.name.hostname |
|
policies[].policy.extAuthz.(any)(1)service.port |
|
policies[].policy.extAuthz.(any)(1)host |
Hostname or IP address |
policies[].policy.extAuthz.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
policies[].policy.extAuthz.(any)protocol |
The ext_authz protocol to use. Unless you need to integrate with an HTTP-only server, gRPC is recommended. |
policies[].policy.extAuthz.(any)protocol.(1)grpc |
|
policies[].policy.extAuthz.(any)protocol.(1)grpc.context |
Additional context to send to the authorization service. This maps to the context_extensions field of the request, and only allows static values. |
policies[].policy.extAuthz.(any)protocol.(1)grpc.metadata |
Additional metadata to send to the authorization service. This maps to the metadata_context.filter_metadata field of the request, and allows dynamic CEL expressions.If unset, by default the envoy.filters.http.jwt_authn key is set if the JWT policy is used as well, for compatibility. |
policies[].policy.extAuthz.(any)protocol.(1)http |
|
policies[].policy.extAuthz.(any)protocol.(1)http.path |
|
policies[].policy.extAuthz.(any)protocol.(1)http.redirect |
When using the HTTP protocol, and the server returns unauthorized, redirect to the URL resolved by the provided expression rather than directly returning the error. |
policies[].policy.extAuthz.(any)protocol.(1)http.includeResponseHeaders |
Specific headers from the authorization response will be copied into the request to the backend. |
policies[].policy.extAuthz.(any)protocol.(1)http.addRequestHeaders |
Specific headers to add in the authorization request (empty = all headers), based on the expression |
policies[].policy.extAuthz.(any)protocol.(1)http.metadata |
Metadata to include under the extauthz variable, based on the authorization response. |
policies[].policy.extAuthz.(any)failureMode |
Behavior when the authorization service is unavailable or returns an error |
policies[].policy.extAuthz.(any)failureMode.(1)denyWithStatus |
|
policies[].policy.extAuthz.(any)includeRequestHeaders |
Specific headers to include in the authorization request. If unset, the gRPC protocol sends all request headers. The HTTP protocol sends only ‘Authorization’. |
policies[].policy.extAuthz.(any)includeRequestBody |
Options for including the request body in the authorization request |
policies[].policy.extAuthz.(any)includeRequestBody.maxRequestBytes |
Maximum size of request body to buffer (default: 8192) |
policies[].policy.extAuthz.(any)includeRequestBody.allowPartialMessage |
If true, send partial body when max_request_bytes is reached |
policies[].policy.extAuthz.(any)includeRequestBody.packAsBytes |
If true, pack body as raw bytes in gRPC |
policies[].policy.extAuthz.(any)timeout |
Timeout for the authorization request (default: 200ms) |
policies[].policy.extProc |
Extend agentgateway with an external processor |
policies[].policy.extProc.(any)(1)service |
|
policies[].policy.extProc.(any)(1)service.name |
|
policies[].policy.extProc.(any)(1)service.name.namespace |
|
policies[].policy.extProc.(any)(1)service.name.hostname |
|
policies[].policy.extProc.(any)(1)service.port |
|
policies[].policy.extProc.(any)(1)host |
Hostname or IP address |
policies[].policy.extProc.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
policies[].policy.extProc.(any)failureMode |
Behavior when the ext_proc service is unavailable or returns an error |
policies[].policy.extProc.(any)metadataContext |
Additional metadata to send to the external processing service. Maps to the metadata_context.filter_metadata field in ProcessingRequest, and allows dynamic CEL expressions. |
policies[].policy.extProc.(any)requestAttributes |
Maps to the request attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
policies[].policy.extProc.(any)responseAttributes |
Maps to the response attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
policies[].policy.transformations |
Modify requests and responses |
policies[].policy.transformations.request |
|
policies[].policy.transformations.request.add |
|
policies[].policy.transformations.request.set |
|
policies[].policy.transformations.request.remove |
|
policies[].policy.transformations.request.body |
|
policies[].policy.transformations.response |
|
policies[].policy.transformations.response.add |
|
policies[].policy.transformations.response.set |
|
policies[].policy.transformations.response.remove |
|
policies[].policy.transformations.response.body |
|
policies[].policy.csrf |
Handle CSRF protection by validating request origins against configured allowed origins. |
policies[].policy.csrf.additionalOrigins |
|
policies[].policy.timeout |
Timeout requests that exceed the configured duration. |
policies[].policy.timeout.requestTimeout |
|
policies[].policy.timeout.backendRequestTimeout |
|
policies[].policy.retry |
Retry matching requests. |
policies[].policy.retry.attempts |
|
policies[].policy.retry.backoff |
|
policies[].policy.retry.codes |
|
workloads |
|
services |
|
backends |
|
backends[].name |
|
backends[].host |
|
backends[].policies |
|
backends[].policies.requestHeaderModifier |
Headers to be modified in the request. |
backends[].policies.requestHeaderModifier.add |
|
backends[].policies.requestHeaderModifier.set |
|
backends[].policies.requestHeaderModifier.remove |
|
backends[].policies.responseHeaderModifier |
Headers to be modified in the response. |
backends[].policies.responseHeaderModifier.add |
|
backends[].policies.responseHeaderModifier.set |
|
backends[].policies.responseHeaderModifier.remove |
|
backends[].policies.requestRedirect |
Directly respond to the request with a redirect. |
backends[].policies.requestRedirect.scheme |
|
backends[].policies.requestRedirect.authority |
|
backends[].policies.requestRedirect.authority.(any)(1)full |
|
backends[].policies.requestRedirect.authority.(any)(1)host |
|
backends[].policies.requestRedirect.authority.(any)(1)port |
|
backends[].policies.requestRedirect.path |
|
backends[].policies.requestRedirect.path.(any)(1)full |
|
backends[].policies.requestRedirect.path.(any)(1)prefix |
|
backends[].policies.requestRedirect.status |
|
backends[].policies.mcpAuthorization |
Authorization policies for MCP access. |
backends[].policies.mcpAuthorization.rules |
|
backends[].policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
backends[].policies.ai |
Mark this as LLM traffic to enable LLM processing. |
backends[].policies.ai.promptGuard |
|
backends[].policies.ai.promptGuard.request |
|
backends[].policies.ai.promptGuard.request[].(1)regex |
|
backends[].policies.ai.promptGuard.request[].(1)regex.action |
|
backends[].policies.ai.promptGuard.request[].(1)regex.rules |
|
backends[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
backends[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
backends[].policies.ai.promptGuard.request[].(1)webhook |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.target |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
backends[].policies.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
backends[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
backends[].policies.ai.promptGuard.request[].(1)openAIModeration |
|
backends[].policies.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
backends[].policies.ai.promptGuard.request[].(1)openAIModeration.policies |
|
backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
backends[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
backends[].policies.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
backends[].policies.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
backends[].policies.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
backends[].policies.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
backends[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
backends[].policies.ai.promptGuard.request[].rejection |
|
backends[].policies.ai.promptGuard.request[].rejection.body |
|
backends[].policies.ai.promptGuard.request[].rejection.status |
|
backends[].policies.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
backends[].policies.ai.promptGuard.request[].rejection.headers.add |
|
backends[].policies.ai.promptGuard.request[].rejection.headers.set |
|
backends[].policies.ai.promptGuard.request[].rejection.headers.remove |
|
backends[].policies.ai.promptGuard.response |
|
backends[].policies.ai.promptGuard.response[].(1)regex |
|
backends[].policies.ai.promptGuard.response[].(1)regex.action |
|
backends[].policies.ai.promptGuard.response[].(1)regex.rules |
|
backends[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
backends[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
backends[].policies.ai.promptGuard.response[].(1)webhook |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.target |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
backends[].policies.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
backends[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
backends[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
backends[].policies.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
backends[].policies.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
backends[].policies.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
backends[].policies.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
backends[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
backends[].policies.ai.promptGuard.response[].rejection |
|
backends[].policies.ai.promptGuard.response[].rejection.body |
|
backends[].policies.ai.promptGuard.response[].rejection.status |
|
backends[].policies.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
backends[].policies.ai.promptGuard.response[].rejection.headers.add |
|
backends[].policies.ai.promptGuard.response[].rejection.headers.set |
|
backends[].policies.ai.promptGuard.response[].rejection.headers.remove |
|
backends[].policies.ai.defaults |
|
backends[].policies.ai.overrides |
|
backends[].policies.ai.transformations |
|
backends[].policies.ai.prompts |
|
backends[].policies.ai.prompts.append |
|
backends[].policies.ai.prompts.append[].role |
|
backends[].policies.ai.prompts.append[].content |
|
backends[].policies.ai.prompts.prepend |
|
backends[].policies.ai.prompts.prepend[].role |
|
backends[].policies.ai.prompts.prepend[].content |
|
backends[].policies.ai.modelAliases |
|
backends[].policies.ai.promptCaching |
|
backends[].policies.ai.promptCaching.cacheSystem |
|
backends[].policies.ai.promptCaching.cacheMessages |
|
backends[].policies.ai.promptCaching.cacheTools |
|
backends[].policies.ai.promptCaching.minTokens |
|
backends[].policies.ai.routes |
|
backends[].policies.backendTLS |
Send TLS to the backend. |
backends[].policies.backendTLS.cert |
|
backends[].policies.backendTLS.key |
|
backends[].policies.backendTLS.root |
|
backends[].policies.backendTLS.hostname |
|
backends[].policies.backendTLS.insecure |
|
backends[].policies.backendTLS.insecureHost |
|
backends[].policies.backendTLS.alpn |
|
backends[].policies.backendTLS.subjectAltNames |
|
backends[].policies.backendAuth |
Authenticate to the backend. |
backends[].policies.backendAuth.(any)(1)passthrough |
|
backends[].policies.backendAuth.(any)(1)key |
|
backends[].policies.backendAuth.(any)(1)key.(any)file |
|
backends[].policies.backendAuth.(any)(1)gcp |
|
backends[].policies.backendAuth.(any)(1)gcp.(any)type |
|
backends[].policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
backends[].policies.backendAuth.(any)(1)gcp.(any)type |
|
backends[].policies.backendAuth.(any)(1)aws |
|
backends[].policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
backends[].policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
backends[].policies.backendAuth.(any)(1)aws.(any)region |
|
backends[].policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
backends[].policies.backendAuth.(any)(1)azure |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
backends[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
backends[].policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
backends[].policies.http |
Specify HTTP settings for the backend |
backends[].policies.http.version |
|
backends[].policies.http.requestTimeout |
|
backends[].policies.tcp |
Specify TCP settings for the backend |
backends[].policies.tcp.keepalives |
|
backends[].policies.tcp.keepalives.enabled |
|
backends[].policies.tcp.keepalives.time |
|
backends[].policies.tcp.keepalives.interval |
|
backends[].policies.tcp.keepalives.retries |
|
backends[].policies.tcp.connectTimeout |
|
backends[].policies.tcp.connectTimeout.secs |
|
backends[].policies.tcp.connectTimeout.nanos |
|
llm |
|
llm.models |
models defines the set of models that can be served by this gateway. The model name refers to the model in the users request that is matched; the model sent to the actual LLM can be overridden on a per-model basis. |
llm.models[].name |
name is the name of the model we are matching from a users request. If params.model is set, that will be used in the request to the LLM provider. If not, the incoming model is used. |
llm.models[].params |
params customizes parameters for the outgoing request |
llm.models[].params.model |
The model to send to the provider. If unset, the same model will be used from the request. |
llm.models[].params.apiKey |
An API key to attach to the request. If unset this will be automatically detected from the environment. |
llm.models[].params.awsRegion |
|
llm.models[].params.vertexRegion |
|
llm.models[].params.vertexProject |
|
llm.models[].params.azureHost |
For Azure: the host of the deployment |
llm.models[].params.azureApiVersion |
For Azure: the API version to use |
llm.models[].provider |
provider of the LLM we are connecting too |
llm.models[].defaults |
defaults allows setting default values for the request. If these are not present in the request body, they will be set. To override even when set, use overrides. |
llm.models[].overrides |
overrides allows setting values for the request, overriding any existing values |
llm.models[].transformation |
transformation allows setting values from CEL expressions for the request, overriding any existing values. |
llm.models[].requestHeaders |
requestHeaders modifies headers in requests to the LLM provider. |
llm.models[].requestHeaders.add |
|
llm.models[].requestHeaders.set |
|
llm.models[].requestHeaders.remove |
|
llm.models[].guardrails |
guardrails to apply to the request or response |
llm.models[].guardrails.request |
|
llm.models[].guardrails.request[].(1)regex |
|
llm.models[].guardrails.request[].(1)regex.action |
|
llm.models[].guardrails.request[].(1)regex.rules |
|
llm.models[].guardrails.request[].(1)regex.rules[].(any)builtin |
|
llm.models[].guardrails.request[].(1)regex.rules[].(any)pattern |
|
llm.models[].guardrails.request[].(1)webhook |
|
llm.models[].guardrails.request[].(1)webhook.target |
|
llm.models[].guardrails.request[].(1)webhook.target.(1)service |
|
llm.models[].guardrails.request[].(1)webhook.target.(1)service.name |
|
llm.models[].guardrails.request[].(1)webhook.target.(1)service.name.namespace |
|
llm.models[].guardrails.request[].(1)webhook.target.(1)service.name.hostname |
|
llm.models[].guardrails.request[].(1)webhook.target.(1)service.port |
|
llm.models[].guardrails.request[].(1)webhook.target.(1)host |
Hostname or IP address |
llm.models[].guardrails.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
llm.models[].guardrails.request[].(1)webhook.forwardHeaderMatches |
|
llm.models[].guardrails.request[].(1)webhook.forwardHeaderMatches[].name |
|
llm.models[].guardrails.request[].(1)webhook.forwardHeaderMatches[].value |
|
llm.models[].guardrails.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
llm.models[].guardrails.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
llm.models[].guardrails.request[].(1)openAIModeration |
|
llm.models[].guardrails.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
llm.models[].guardrails.request[].(1)openAIModeration.policies |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestHeaderModifier |
Headers to be modified in the request. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestHeaderModifier.add |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestHeaderModifier.set |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestHeaderModifier.remove |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.responseHeaderModifier |
Headers to be modified in the response. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.responseHeaderModifier.add |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.responseHeaderModifier.set |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.responseHeaderModifier.remove |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect |
Directly respond to the request with a redirect. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.scheme |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.authority |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)full |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)host |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)port |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.path |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.path.(any)(1)full |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.path.(any)(1)prefix |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.requestRedirect.status |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.mcpAuthorization |
Authorization policies for MCP access. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.mcpAuthorization.rules |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.promptGuard |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.defaults |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.overrides |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.transformations |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.prompts |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.prompts.append |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.prompts.append[].role |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.prompts.append[].content |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.prompts.prepend |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.prompts.prepend[].role |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.prompts.prepend[].content |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.modelAliases |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.promptCaching |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.promptCaching.cacheSystem |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.promptCaching.cacheMessages |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.promptCaching.cacheTools |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.promptCaching.minTokens |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.ai.routes |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS |
Send TLS to the backend. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS.cert |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS.key |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS.root |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS.hostname |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS.insecure |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS.insecureHost |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS.alpn |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendTLS.subjectAltNames |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth |
Authenticate to the backend. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)passthrough |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)key |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)key.(any)file |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)region |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.http |
Specify HTTP settings for the backend |
llm.models[].guardrails.request[].(1)openAIModeration.policies.http.version |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.http.requestTimeout |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp |
Specify TCP settings for the backend |
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp.keepalives |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp.keepalives.enabled |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp.keepalives.time |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp.keepalives.interval |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp.keepalives.retries |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp.connectTimeout |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp.connectTimeout.secs |
|
llm.models[].guardrails.request[].(1)openAIModeration.policies.tcp.connectTimeout.nanos |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
llm.models[].guardrails.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
llm.models[].guardrails.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestHeaderModifier |
Headers to be modified in the request. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.add |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.set |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.remove |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.responseHeaderModifier |
Headers to be modified in the response. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.add |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.set |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.remove |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect |
Directly respond to the request with a redirect. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.scheme |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.authority |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)full |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)host |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)port |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.path |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)full |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)prefix |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.requestRedirect.status |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.mcpAuthorization |
Authorization policies for MCP access. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.mcpAuthorization.rules |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.promptGuard |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.defaults |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.overrides |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.transformations |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.prompts |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.prompts.append |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.prompts.append[].role |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.prompts.append[].content |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.prompts.prepend |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.prompts.prepend[].role |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.prompts.prepend[].content |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.modelAliases |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.promptCaching |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.promptCaching.cacheSystem |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.promptCaching.cacheMessages |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.promptCaching.cacheTools |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.promptCaching.minTokens |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.ai.routes |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS |
Send TLS to the backend. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS.cert |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS.key |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS.root |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS.hostname |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS.insecure |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS.insecureHost |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS.alpn |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendTLS.subjectAltNames |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth |
Authenticate to the backend. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)passthrough |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key.(any)file |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)region |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.http |
Specify HTTP settings for the backend |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.http.version |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.http.requestTimeout |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp |
Specify TCP settings for the backend |
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp.keepalives |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp.keepalives.enabled |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp.keepalives.time |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp.keepalives.interval |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp.keepalives.retries |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout.secs |
|
llm.models[].guardrails.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout.nanos |
|
llm.models[].guardrails.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
llm.models[].guardrails.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
llm.models[].guardrails.request[].(1)googleModelArmor.projectId |
The GCP project ID |
llm.models[].guardrails.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
llm.models[].guardrails.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestHeaderModifier |
Headers to be modified in the request. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestHeaderModifier.add |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestHeaderModifier.set |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestHeaderModifier.remove |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.responseHeaderModifier |
Headers to be modified in the response. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.responseHeaderModifier.add |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.responseHeaderModifier.set |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.responseHeaderModifier.remove |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect |
Directly respond to the request with a redirect. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.scheme |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.authority |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)full |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)host |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)port |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.path |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)full |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)prefix |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.requestRedirect.status |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.mcpAuthorization |
Authorization policies for MCP access. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.mcpAuthorization.rules |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.promptGuard |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.defaults |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.overrides |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.transformations |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.prompts |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.prompts.append |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.prompts.append[].role |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.prompts.append[].content |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.prompts.prepend |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.prompts.prepend[].role |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.prompts.prepend[].content |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.modelAliases |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.promptCaching |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.promptCaching.cacheSystem |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.promptCaching.cacheMessages |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.promptCaching.cacheTools |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.promptCaching.minTokens |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.ai.routes |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS |
Send TLS to the backend. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS.cert |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS.key |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS.root |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS.hostname |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS.insecure |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS.insecureHost |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS.alpn |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendTLS.subjectAltNames |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth |
Authenticate to the backend. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)passthrough |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)key |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)key.(any)file |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)region |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.http |
Specify HTTP settings for the backend |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.http.version |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.http.requestTimeout |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp |
Specify TCP settings for the backend |
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp.keepalives |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp.keepalives.enabled |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp.keepalives.time |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp.keepalives.interval |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp.keepalives.retries |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp.connectTimeout |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp.connectTimeout.secs |
|
llm.models[].guardrails.request[].(1)googleModelArmor.policies.tcp.connectTimeout.nanos |
|
llm.models[].guardrails.request[].rejection |
|
llm.models[].guardrails.request[].rejection.body |
|
llm.models[].guardrails.request[].rejection.status |
|
llm.models[].guardrails.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
llm.models[].guardrails.request[].rejection.headers.add |
|
llm.models[].guardrails.request[].rejection.headers.set |
|
llm.models[].guardrails.request[].rejection.headers.remove |
|
llm.models[].guardrails.response |
|
llm.models[].guardrails.response[].(1)regex |
|
llm.models[].guardrails.response[].(1)regex.action |
|
llm.models[].guardrails.response[].(1)regex.rules |
|
llm.models[].guardrails.response[].(1)regex.rules[].(any)builtin |
|
llm.models[].guardrails.response[].(1)regex.rules[].(any)pattern |
|
llm.models[].guardrails.response[].(1)webhook |
|
llm.models[].guardrails.response[].(1)webhook.target |
|
llm.models[].guardrails.response[].(1)webhook.target.(1)service |
|
llm.models[].guardrails.response[].(1)webhook.target.(1)service.name |
|
llm.models[].guardrails.response[].(1)webhook.target.(1)service.name.namespace |
|
llm.models[].guardrails.response[].(1)webhook.target.(1)service.name.hostname |
|
llm.models[].guardrails.response[].(1)webhook.target.(1)service.port |
|
llm.models[].guardrails.response[].(1)webhook.target.(1)host |
Hostname or IP address |
llm.models[].guardrails.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
llm.models[].guardrails.response[].(1)webhook.forwardHeaderMatches |
|
llm.models[].guardrails.response[].(1)webhook.forwardHeaderMatches[].name |
|
llm.models[].guardrails.response[].(1)webhook.forwardHeaderMatches[].value |
|
llm.models[].guardrails.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
llm.models[].guardrails.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
llm.models[].guardrails.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
llm.models[].guardrails.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestHeaderModifier |
Headers to be modified in the request. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.add |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.set |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.remove |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.responseHeaderModifier |
Headers to be modified in the response. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.add |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.set |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.remove |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect |
Directly respond to the request with a redirect. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.scheme |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.authority |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)full |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)host |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)port |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.path |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)full |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)prefix |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.requestRedirect.status |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.mcpAuthorization |
Authorization policies for MCP access. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.mcpAuthorization.rules |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.promptGuard |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.defaults |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.overrides |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.transformations |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.prompts |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.prompts.append |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.prompts.append[].role |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.prompts.append[].content |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.prompts.prepend |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.prompts.prepend[].role |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.prompts.prepend[].content |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.modelAliases |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.promptCaching |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.promptCaching.cacheSystem |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.promptCaching.cacheMessages |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.promptCaching.cacheTools |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.promptCaching.minTokens |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.ai.routes |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS |
Send TLS to the backend. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS.cert |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS.key |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS.root |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS.hostname |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS.insecure |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS.insecureHost |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS.alpn |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendTLS.subjectAltNames |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth |
Authenticate to the backend. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)passthrough |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key.(any)file |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)region |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.http |
Specify HTTP settings for the backend |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.http.version |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.http.requestTimeout |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp |
Specify TCP settings for the backend |
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp.keepalives |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp.keepalives.enabled |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp.keepalives.time |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp.keepalives.interval |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp.keepalives.retries |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout.secs |
|
llm.models[].guardrails.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout.nanos |
|
llm.models[].guardrails.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
llm.models[].guardrails.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
llm.models[].guardrails.response[].(1)googleModelArmor.projectId |
The GCP project ID |
llm.models[].guardrails.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
llm.models[].guardrails.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestHeaderModifier |
Headers to be modified in the request. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestHeaderModifier.add |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestHeaderModifier.set |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestHeaderModifier.remove |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.responseHeaderModifier |
Headers to be modified in the response. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.responseHeaderModifier.add |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.responseHeaderModifier.set |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.responseHeaderModifier.remove |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect |
Directly respond to the request with a redirect. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.scheme |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.authority |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)full |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)host |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)port |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.path |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)full |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)prefix |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.requestRedirect.status |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.mcpAuthorization |
Authorization policies for MCP access. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.mcpAuthorization.rules |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.promptGuard |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.defaults |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.overrides |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.transformations |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.prompts |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.prompts.append |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.prompts.append[].role |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.prompts.append[].content |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.prompts.prepend |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.prompts.prepend[].role |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.prompts.prepend[].content |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.modelAliases |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.promptCaching |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.promptCaching.cacheSystem |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.promptCaching.cacheMessages |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.promptCaching.cacheTools |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.promptCaching.minTokens |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.ai.routes |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS |
Send TLS to the backend. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS.cert |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS.key |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS.root |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS.hostname |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS.insecure |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS.insecureHost |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS.alpn |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendTLS.subjectAltNames |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth |
Authenticate to the backend. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)passthrough |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)key |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)key.(any)file |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)region |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.http |
Specify HTTP settings for the backend |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.http.version |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.http.requestTimeout |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp |
Specify TCP settings for the backend |
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp.keepalives |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp.keepalives.enabled |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp.keepalives.time |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp.keepalives.interval |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp.keepalives.retries |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp.connectTimeout |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp.connectTimeout.secs |
|
llm.models[].guardrails.response[].(1)googleModelArmor.policies.tcp.connectTimeout.nanos |
|
llm.models[].guardrails.response[].rejection |
|
llm.models[].guardrails.response[].rejection.body |
|
llm.models[].guardrails.response[].rejection.status |
|
llm.models[].guardrails.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
llm.models[].guardrails.response[].rejection.headers.add |
|
llm.models[].guardrails.response[].rejection.headers.set |
|
llm.models[].guardrails.response[].rejection.headers.remove |
|
llm.models[].matches |
matches specifies the conditions under which this model should be used in addition to matching the model name. |
llm.models[].matches[].headers |
|
llm.models[].matches[].headers[].name |
|
llm.models[].matches[].headers[].value |
|
llm.models[].matches[].headers[].value.(1)exact |
|
llm.models[].matches[].headers[].value.(1)regex |
|
llm.policies |
policies defines policies for handling incoming requests, before a model is selected |
llm.policies.jwtAuth |
Authenticate incoming JWT requests. |
llm.policies.jwtAuth.(any)(any)mode |
|
llm.policies.jwtAuth.(any)(any)providers |
|
llm.policies.jwtAuth.(any)(any)providers[].issuer |
|
llm.policies.jwtAuth.(any)(any)providers[].audiences |
|
llm.policies.jwtAuth.(any)(any)providers[].jwks |
|
llm.policies.jwtAuth.(any)(any)providers[].jwks.(any)file |
|
llm.policies.jwtAuth.(any)(any)providers[].jwks.(any)url |
|
llm.policies.jwtAuth.(any)(any)mode |
|
llm.policies.jwtAuth.(any)(any)issuer |
|
llm.policies.jwtAuth.(any)(any)audiences |
|
llm.policies.jwtAuth.(any)(any)jwks |
|
llm.policies.jwtAuth.(any)(any)jwks.(any)file |
|
llm.policies.jwtAuth.(any)(any)jwks.(any)url |
|
llm.policies.extAuthz |
Authenticate incoming requests by calling an external authorization server. |
llm.policies.extAuthz.(any)(1)service |
|
llm.policies.extAuthz.(any)(1)service.name |
|
llm.policies.extAuthz.(any)(1)service.name.namespace |
|
llm.policies.extAuthz.(any)(1)service.name.hostname |
|
llm.policies.extAuthz.(any)(1)service.port |
|
llm.policies.extAuthz.(any)(1)host |
Hostname or IP address |
llm.policies.extAuthz.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
llm.policies.extAuthz.(any)protocol |
The ext_authz protocol to use. Unless you need to integrate with an HTTP-only server, gRPC is recommended. |
llm.policies.extAuthz.(any)protocol.(1)grpc |
|
llm.policies.extAuthz.(any)protocol.(1)grpc.context |
Additional context to send to the authorization service. This maps to the context_extensions field of the request, and only allows static values. |
llm.policies.extAuthz.(any)protocol.(1)grpc.metadata |
Additional metadata to send to the authorization service. This maps to the metadata_context.filter_metadata field of the request, and allows dynamic CEL expressions.If unset, by default the envoy.filters.http.jwt_authn key is set if the JWT policy is used as well, for compatibility. |
llm.policies.extAuthz.(any)protocol.(1)http |
|
llm.policies.extAuthz.(any)protocol.(1)http.path |
|
llm.policies.extAuthz.(any)protocol.(1)http.redirect |
When using the HTTP protocol, and the server returns unauthorized, redirect to the URL resolved by the provided expression rather than directly returning the error. |
llm.policies.extAuthz.(any)protocol.(1)http.includeResponseHeaders |
Specific headers from the authorization response will be copied into the request to the backend. |
llm.policies.extAuthz.(any)protocol.(1)http.addRequestHeaders |
Specific headers to add in the authorization request (empty = all headers), based on the expression |
llm.policies.extAuthz.(any)protocol.(1)http.metadata |
Metadata to include under the extauthz variable, based on the authorization response. |
llm.policies.extAuthz.(any)failureMode |
Behavior when the authorization service is unavailable or returns an error |
llm.policies.extAuthz.(any)failureMode.(1)denyWithStatus |
|
llm.policies.extAuthz.(any)includeRequestHeaders |
Specific headers to include in the authorization request. If unset, the gRPC protocol sends all request headers. The HTTP protocol sends only ‘Authorization’. |
llm.policies.extAuthz.(any)includeRequestBody |
Options for including the request body in the authorization request |
llm.policies.extAuthz.(any)includeRequestBody.maxRequestBytes |
Maximum size of request body to buffer (default: 8192) |
llm.policies.extAuthz.(any)includeRequestBody.allowPartialMessage |
If true, send partial body when max_request_bytes is reached |
llm.policies.extAuthz.(any)includeRequestBody.packAsBytes |
If true, pack body as raw bytes in gRPC |
llm.policies.extAuthz.(any)timeout |
Timeout for the authorization request (default: 200ms) |
llm.policies.extProc |
Extend agentgateway with an external processor |
llm.policies.extProc.(any)(1)service |
|
llm.policies.extProc.(any)(1)service.name |
|
llm.policies.extProc.(any)(1)service.name.namespace |
|
llm.policies.extProc.(any)(1)service.name.hostname |
|
llm.policies.extProc.(any)(1)service.port |
|
llm.policies.extProc.(any)(1)host |
Hostname or IP address |
llm.policies.extProc.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
llm.policies.extProc.(any)failureMode |
Behavior when the ext_proc service is unavailable or returns an error |
llm.policies.extProc.(any)metadataContext |
Additional metadata to send to the external processing service. Maps to the metadata_context.filter_metadata field in ProcessingRequest, and allows dynamic CEL expressions. |
llm.policies.extProc.(any)requestAttributes |
Maps to the request attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
llm.policies.extProc.(any)responseAttributes |
Maps to the response attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
llm.policies.transformations |
Modify requests and responses |
llm.policies.transformations.request |
|
llm.policies.transformations.request.add |
|
llm.policies.transformations.request.set |
|
llm.policies.transformations.request.remove |
|
llm.policies.transformations.request.body |
|
llm.policies.transformations.response |
|
llm.policies.transformations.response.add |
|
llm.policies.transformations.response.set |
|
llm.policies.transformations.response.remove |
|
llm.policies.transformations.response.body |
|
llm.policies.basicAuth |
Authenticate incoming requests using Basic Authentication with htpasswd. |
llm.policies.basicAuth.htpasswd |
.htpasswd file contents/reference |
llm.policies.basicAuth.htpasswd.(any)file |
|
llm.policies.basicAuth.realm |
Realm name for the WWW-Authenticate header |
llm.policies.basicAuth.mode |
Validation mode for basic authentication |
llm.policies.apiKey |
Authenticate incoming requests using API Keys |
llm.policies.apiKey.keys |
List of API keys |
llm.policies.apiKey.keys[].key |
|
llm.policies.apiKey.keys[].metadata |
|
llm.policies.apiKey.mode |
Validation mode for API keys |
llm.policies.authorization |
Authorization policies for HTTP access. |
llm.policies.authorization.rules |
|
mcp |
|
mcp.port |
|
mcp.targets |
|
mcp.targets[].(1)sse |
|
mcp.targets[].(1)sse.host |
|
mcp.targets[].(1)sse.port |
|
mcp.targets[].(1)sse.path |
|
mcp.targets[].(1)mcp |
|
mcp.targets[].(1)mcp.host |
|
mcp.targets[].(1)mcp.port |
|
mcp.targets[].(1)mcp.path |
|
mcp.targets[].(1)stdio |
|
mcp.targets[].(1)stdio.cmd |
|
mcp.targets[].(1)stdio.args |
|
mcp.targets[].(1)stdio.env |
|
mcp.targets[].(1)openapi |
|
mcp.targets[].(1)openapi.host |
|
mcp.targets[].(1)openapi.port |
|
mcp.targets[].(1)openapi.path |
|
mcp.targets[].(1)openapi.schema |
|
mcp.targets[].name |
|
mcp.targets[].policies |
|
mcp.targets[].policies.requestHeaderModifier |
Headers to be modified in the request. |
mcp.targets[].policies.requestHeaderModifier.add |
|
mcp.targets[].policies.requestHeaderModifier.set |
|
mcp.targets[].policies.requestHeaderModifier.remove |
|
mcp.targets[].policies.responseHeaderModifier |
Headers to be modified in the response. |
mcp.targets[].policies.responseHeaderModifier.add |
|
mcp.targets[].policies.responseHeaderModifier.set |
|
mcp.targets[].policies.responseHeaderModifier.remove |
|
mcp.targets[].policies.requestRedirect |
Directly respond to the request with a redirect. |
mcp.targets[].policies.requestRedirect.scheme |
|
mcp.targets[].policies.requestRedirect.authority |
|
mcp.targets[].policies.requestRedirect.authority.(any)(1)full |
|
mcp.targets[].policies.requestRedirect.authority.(any)(1)host |
|
mcp.targets[].policies.requestRedirect.authority.(any)(1)port |
|
mcp.targets[].policies.requestRedirect.path |
|
mcp.targets[].policies.requestRedirect.path.(any)(1)full |
|
mcp.targets[].policies.requestRedirect.path.(any)(1)prefix |
|
mcp.targets[].policies.requestRedirect.status |
|
mcp.targets[].policies.mcpAuthorization |
Authorization policies for MCP access. |
mcp.targets[].policies.mcpAuthorization.rules |
|
mcp.targets[].policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
mcp.targets[].policies.ai |
Mark this as LLM traffic to enable LLM processing. |
mcp.targets[].policies.ai.promptGuard |
|
mcp.targets[].policies.ai.promptGuard.request |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)regex |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)regex.action |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)regex.rules |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)openAIModeration |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
mcp.targets[].policies.ai.promptGuard.request[].(1)openAIModeration.policies |
|
mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
mcp.targets[].policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
mcp.targets[].policies.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
mcp.targets[].policies.ai.promptGuard.request[].rejection |
|
mcp.targets[].policies.ai.promptGuard.request[].rejection.body |
|
mcp.targets[].policies.ai.promptGuard.request[].rejection.status |
|
mcp.targets[].policies.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
mcp.targets[].policies.ai.promptGuard.request[].rejection.headers.add |
|
mcp.targets[].policies.ai.promptGuard.request[].rejection.headers.set |
|
mcp.targets[].policies.ai.promptGuard.request[].rejection.headers.remove |
|
mcp.targets[].policies.ai.promptGuard.response |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)regex |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)regex.action |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)regex.rules |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
mcp.targets[].policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
mcp.targets[].policies.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
mcp.targets[].policies.ai.promptGuard.response[].rejection |
|
mcp.targets[].policies.ai.promptGuard.response[].rejection.body |
|
mcp.targets[].policies.ai.promptGuard.response[].rejection.status |
|
mcp.targets[].policies.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
mcp.targets[].policies.ai.promptGuard.response[].rejection.headers.add |
|
mcp.targets[].policies.ai.promptGuard.response[].rejection.headers.set |
|
mcp.targets[].policies.ai.promptGuard.response[].rejection.headers.remove |
|
mcp.targets[].policies.ai.defaults |
|
mcp.targets[].policies.ai.overrides |
|
mcp.targets[].policies.ai.transformations |
|
mcp.targets[].policies.ai.prompts |
|
mcp.targets[].policies.ai.prompts.append |
|
mcp.targets[].policies.ai.prompts.append[].role |
|
mcp.targets[].policies.ai.prompts.append[].content |
|
mcp.targets[].policies.ai.prompts.prepend |
|
mcp.targets[].policies.ai.prompts.prepend[].role |
|
mcp.targets[].policies.ai.prompts.prepend[].content |
|
mcp.targets[].policies.ai.modelAliases |
|
mcp.targets[].policies.ai.promptCaching |
|
mcp.targets[].policies.ai.promptCaching.cacheSystem |
|
mcp.targets[].policies.ai.promptCaching.cacheMessages |
|
mcp.targets[].policies.ai.promptCaching.cacheTools |
|
mcp.targets[].policies.ai.promptCaching.minTokens |
|
mcp.targets[].policies.ai.routes |
|
mcp.targets[].policies.backendTLS |
Send TLS to the backend. |
mcp.targets[].policies.backendTLS.cert |
|
mcp.targets[].policies.backendTLS.key |
|
mcp.targets[].policies.backendTLS.root |
|
mcp.targets[].policies.backendTLS.hostname |
|
mcp.targets[].policies.backendTLS.insecure |
|
mcp.targets[].policies.backendTLS.insecureHost |
|
mcp.targets[].policies.backendTLS.alpn |
|
mcp.targets[].policies.backendTLS.subjectAltNames |
|
mcp.targets[].policies.backendAuth |
Authenticate to the backend. |
mcp.targets[].policies.backendAuth.(any)(1)passthrough |
|
mcp.targets[].policies.backendAuth.(any)(1)key |
|
mcp.targets[].policies.backendAuth.(any)(1)key.(any)file |
|
mcp.targets[].policies.backendAuth.(any)(1)gcp |
|
mcp.targets[].policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.targets[].policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
mcp.targets[].policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.targets[].policies.backendAuth.(any)(1)aws |
|
mcp.targets[].policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
mcp.targets[].policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
mcp.targets[].policies.backendAuth.(any)(1)aws.(any)region |
|
mcp.targets[].policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
mcp.targets[].policies.backendAuth.(any)(1)azure |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
mcp.targets[].policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
mcp.targets[].policies.http |
Specify HTTP settings for the backend |
mcp.targets[].policies.http.version |
|
mcp.targets[].policies.http.requestTimeout |
|
mcp.targets[].policies.tcp |
Specify TCP settings for the backend |
mcp.targets[].policies.tcp.keepalives |
|
mcp.targets[].policies.tcp.keepalives.enabled |
|
mcp.targets[].policies.tcp.keepalives.time |
|
mcp.targets[].policies.tcp.keepalives.interval |
|
mcp.targets[].policies.tcp.keepalives.retries |
|
mcp.targets[].policies.tcp.connectTimeout |
|
mcp.targets[].policies.tcp.connectTimeout.secs |
|
mcp.targets[].policies.tcp.connectTimeout.nanos |
|
mcp.statefulMode |
|
mcp.prefixMode |
|
mcp.policies |
|
mcp.policies.requestHeaderModifier |
Headers to be modified in the request. |
mcp.policies.requestHeaderModifier.add |
|
mcp.policies.requestHeaderModifier.set |
|
mcp.policies.requestHeaderModifier.remove |
|
mcp.policies.responseHeaderModifier |
Headers to be modified in the response. |
mcp.policies.responseHeaderModifier.add |
|
mcp.policies.responseHeaderModifier.set |
|
mcp.policies.responseHeaderModifier.remove |
|
mcp.policies.requestRedirect |
Directly respond to the request with a redirect. |
mcp.policies.requestRedirect.scheme |
|
mcp.policies.requestRedirect.authority |
|
mcp.policies.requestRedirect.authority.(any)(1)full |
|
mcp.policies.requestRedirect.authority.(any)(1)host |
|
mcp.policies.requestRedirect.authority.(any)(1)port |
|
mcp.policies.requestRedirect.path |
|
mcp.policies.requestRedirect.path.(any)(1)full |
|
mcp.policies.requestRedirect.path.(any)(1)prefix |
|
mcp.policies.requestRedirect.status |
|
mcp.policies.urlRewrite |
Modify the URL path or authority. |
mcp.policies.urlRewrite.authority |
|
mcp.policies.urlRewrite.authority.(any)(1)full |
|
mcp.policies.urlRewrite.authority.(any)(1)host |
|
mcp.policies.urlRewrite.authority.(any)(1)port |
|
mcp.policies.urlRewrite.path |
|
mcp.policies.urlRewrite.path.(any)(1)full |
|
mcp.policies.urlRewrite.path.(any)(1)prefix |
|
mcp.policies.requestMirror |
Mirror incoming requests to another destination. |
mcp.policies.requestMirror.backend |
|
mcp.policies.requestMirror.backend.(1)service |
|
mcp.policies.requestMirror.backend.(1)service.name |
|
mcp.policies.requestMirror.backend.(1)service.name.namespace |
|
mcp.policies.requestMirror.backend.(1)service.name.hostname |
|
mcp.policies.requestMirror.backend.(1)service.port |
|
mcp.policies.requestMirror.backend.(1)host |
Hostname or IP address |
mcp.policies.requestMirror.backend.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
mcp.policies.requestMirror.percentage |
|
mcp.policies.directResponse |
Directly respond to the request with a static response. |
mcp.policies.directResponse.body |
|
mcp.policies.directResponse.status |
|
mcp.policies.cors |
Handle CORS preflight requests and append configured CORS headers to applicable requests. |
mcp.policies.cors.allowCredentials |
|
mcp.policies.cors.allowHeaders |
|
mcp.policies.cors.allowMethods |
|
mcp.policies.cors.allowOrigins |
|
mcp.policies.cors.exposeHeaders |
|
mcp.policies.cors.maxAge |
|
mcp.policies.mcpAuthorization |
Authorization policies for MCP access. |
mcp.policies.mcpAuthorization.rules |
|
mcp.policies.authorization |
Authorization policies for HTTP access. |
mcp.policies.authorization.rules |
|
mcp.policies.mcpAuthentication |
Authentication for MCP clients. |
mcp.policies.mcpAuthentication.issuer |
|
mcp.policies.mcpAuthentication.audiences |
|
mcp.policies.mcpAuthentication.provider |
|
mcp.policies.mcpAuthentication.provider.(any)(1)auth0 |
|
mcp.policies.mcpAuthentication.provider.(any)(1)keycloak |
|
mcp.policies.mcpAuthentication.resourceMetadata |
|
mcp.policies.mcpAuthentication.jwks |
|
mcp.policies.mcpAuthentication.jwks.(any)file |
|
mcp.policies.mcpAuthentication.jwks.(any)url |
|
mcp.policies.mcpAuthentication.mode |
|
mcp.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
mcp.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
mcp.policies.ai.promptGuard |
|
mcp.policies.ai.promptGuard.request |
|
mcp.policies.ai.promptGuard.request[].(1)regex |
|
mcp.policies.ai.promptGuard.request[].(1)regex.action |
|
mcp.policies.ai.promptGuard.request[].(1)regex.rules |
|
mcp.policies.ai.promptGuard.request[].(1)regex.rules[].(any)builtin |
|
mcp.policies.ai.promptGuard.request[].(1)regex.rules[].(any)pattern |
|
mcp.policies.ai.promptGuard.request[].(1)webhook |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.target |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.target.(1)service |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.namespace |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.target.(1)service.name.hostname |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.target.(1)service.port |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.target.(1)host |
Hostname or IP address |
mcp.policies.ai.promptGuard.request[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
mcp.policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].name |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
mcp.policies.ai.promptGuard.request[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.model |
Model to use. Defaults to omni-moderation-latest |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier |
Headers to be modified in the request. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.add |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.set |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier |
Headers to be modified in the response. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.add |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.set |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.responseHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect |
Directly respond to the request with a redirect. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.scheme |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)full |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)host |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.authority.(any)(1)port |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path.(any)(1)full |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.path.(any)(1)prefix |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.requestRedirect.status |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.mcpAuthorization |
Authorization policies for MCP access. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.mcpAuthorization.rules |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS |
Send TLS to the backend. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.cert |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.key |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.root |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.hostname |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.insecure |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.insecureHost |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.alpn |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendTLS.subjectAltNames |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth |
Authenticate to the backend. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)passthrough |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)key |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)key.(any)file |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)region |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.http |
Specify HTTP settings for the backend |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.http.version |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.http.requestTimeout |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp |
Specify TCP settings for the backend |
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.enabled |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.time |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.interval |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.keepalives.retries |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout.secs |
|
mcp.policies.ai.promptGuard.request[].(1)openAIModeration.policies.tcp.connectTimeout.nanos |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier |
Headers to be modified in the request. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.add |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.set |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier |
Headers to be modified in the response. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.add |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.set |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.responseHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect |
Directly respond to the request with a redirect. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.scheme |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)full |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)host |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)port |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)full |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)prefix |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.requestRedirect.status |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.mcpAuthorization |
Authorization policies for MCP access. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.mcpAuthorization.rules |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS |
Send TLS to the backend. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.cert |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.key |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.root |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.hostname |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.insecure |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.insecureHost |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.alpn |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendTLS.subjectAltNames |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth |
Authenticate to the backend. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)passthrough |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key.(any)file |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)region |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http |
Specify HTTP settings for the backend |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http.version |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.http.requestTimeout |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp |
Specify TCP settings for the backend |
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.enabled |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.time |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.interval |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.keepalives.retries |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout.secs |
|
mcp.policies.ai.promptGuard.request[].(1)bedrockGuardrails.policies.tcp.connectTimeout.nanos |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.projectId |
The GCP project ID |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier |
Headers to be modified in the request. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.add |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.set |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier |
Headers to be modified in the response. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.add |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.set |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.responseHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect |
Directly respond to the request with a redirect. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.scheme |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)full |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)host |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)port |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)full |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)prefix |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.requestRedirect.status |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.mcpAuthorization |
Authorization policies for MCP access. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.mcpAuthorization.rules |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS |
Send TLS to the backend. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.cert |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.key |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.root |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.hostname |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.insecure |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.insecureHost |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.alpn |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendTLS.subjectAltNames |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth |
Authenticate to the backend. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)passthrough |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)key |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)key.(any)file |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)region |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.http |
Specify HTTP settings for the backend |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.http.version |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.http.requestTimeout |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp |
Specify TCP settings for the backend |
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.enabled |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.time |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.interval |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.keepalives.retries |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout.secs |
|
mcp.policies.ai.promptGuard.request[].(1)googleModelArmor.policies.tcp.connectTimeout.nanos |
|
mcp.policies.ai.promptGuard.request[].rejection |
|
mcp.policies.ai.promptGuard.request[].rejection.body |
|
mcp.policies.ai.promptGuard.request[].rejection.status |
|
mcp.policies.ai.promptGuard.request[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
mcp.policies.ai.promptGuard.request[].rejection.headers.add |
|
mcp.policies.ai.promptGuard.request[].rejection.headers.set |
|
mcp.policies.ai.promptGuard.request[].rejection.headers.remove |
|
mcp.policies.ai.promptGuard.response |
|
mcp.policies.ai.promptGuard.response[].(1)regex |
|
mcp.policies.ai.promptGuard.response[].(1)regex.action |
|
mcp.policies.ai.promptGuard.response[].(1)regex.rules |
|
mcp.policies.ai.promptGuard.response[].(1)regex.rules[].(any)builtin |
|
mcp.policies.ai.promptGuard.response[].(1)regex.rules[].(any)pattern |
|
mcp.policies.ai.promptGuard.response[].(1)webhook |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.target |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.target.(1)service |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.namespace |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.target.(1)service.name.hostname |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.target.(1)service.port |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.target.(1)host |
Hostname or IP address |
mcp.policies.ai.promptGuard.response[].(1)webhook.target.(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
mcp.policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].name |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)exact |
|
mcp.policies.ai.promptGuard.response[].(1)webhook.forwardHeaderMatches[].value.(1)regex |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails |
Configuration for AWS Bedrock Guardrails integration. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailIdentifier |
The unique identifier of the guardrail |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.guardrailVersion |
The version of the guardrail |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.region |
AWS region where the guardrail is deployed |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies |
Backend policies for AWS authentication (optional, defaults to implicit AWS auth) |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier |
Headers to be modified in the request. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.add |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.set |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier |
Headers to be modified in the response. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.add |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.set |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.responseHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect |
Directly respond to the request with a redirect. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.scheme |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)full |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)host |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.authority.(any)(1)port |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)full |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.path.(any)(1)prefix |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.requestRedirect.status |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.mcpAuthorization |
Authorization policies for MCP access. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.mcpAuthorization.rules |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS |
Send TLS to the backend. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.cert |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.key |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.root |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.hostname |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.insecure |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.insecureHost |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.alpn |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendTLS.subjectAltNames |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth |
Authenticate to the backend. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)passthrough |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)key.(any)file |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)region |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http |
Specify HTTP settings for the backend |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http.version |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.http.requestTimeout |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp |
Specify TCP settings for the backend |
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.enabled |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.time |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.interval |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.keepalives.retries |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout.secs |
|
mcp.policies.ai.promptGuard.response[].(1)bedrockGuardrails.policies.tcp.connectTimeout.nanos |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor |
Configuration for Google Cloud Model Armor integration. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.templateId |
The template ID for the Model Armor configuration |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.projectId |
The GCP project ID |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.location |
The GCP region (default: us-central1) |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies |
Backend policies for GCP authentication (optional, defaults to implicit GCP auth) |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier |
Headers to be modified in the request. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.add |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.set |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier |
Headers to be modified in the response. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.add |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.set |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.responseHeaderModifier.remove |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect |
Directly respond to the request with a redirect. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.scheme |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)full |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)host |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.authority.(any)(1)port |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)full |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.path.(any)(1)prefix |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.requestRedirect.status |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.mcpAuthorization |
Authorization policies for MCP access. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.mcpAuthorization.rules |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.a2a |
Mark this traffic as A2A to enable A2A processing and telemetry. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.ai |
Mark this as LLM traffic to enable LLM processing. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS |
Send TLS to the backend. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.cert |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.key |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.root |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.hostname |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.insecure |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.insecureHost |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.alpn |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendTLS.subjectAltNames |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth |
Authenticate to the backend. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)passthrough |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)key |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)key.(any)file |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)region |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.http |
Specify HTTP settings for the backend |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.http.version |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.http.requestTimeout |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp |
Specify TCP settings for the backend |
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.enabled |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.time |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.interval |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.keepalives.retries |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout.secs |
|
mcp.policies.ai.promptGuard.response[].(1)googleModelArmor.policies.tcp.connectTimeout.nanos |
|
mcp.policies.ai.promptGuard.response[].rejection |
|
mcp.policies.ai.promptGuard.response[].rejection.body |
|
mcp.policies.ai.promptGuard.response[].rejection.status |
|
mcp.policies.ai.promptGuard.response[].rejection.headers |
Optional headers to add, set, or remove from the rejection response |
mcp.policies.ai.promptGuard.response[].rejection.headers.add |
|
mcp.policies.ai.promptGuard.response[].rejection.headers.set |
|
mcp.policies.ai.promptGuard.response[].rejection.headers.remove |
|
mcp.policies.ai.defaults |
|
mcp.policies.ai.overrides |
|
mcp.policies.ai.transformations |
|
mcp.policies.ai.prompts |
|
mcp.policies.ai.prompts.append |
|
mcp.policies.ai.prompts.append[].role |
|
mcp.policies.ai.prompts.append[].content |
|
mcp.policies.ai.prompts.prepend |
|
mcp.policies.ai.prompts.prepend[].role |
|
mcp.policies.ai.prompts.prepend[].content |
|
mcp.policies.ai.modelAliases |
|
mcp.policies.ai.promptCaching |
|
mcp.policies.ai.promptCaching.cacheSystem |
|
mcp.policies.ai.promptCaching.cacheMessages |
|
mcp.policies.ai.promptCaching.cacheTools |
|
mcp.policies.ai.promptCaching.minTokens |
|
mcp.policies.ai.routes |
|
mcp.policies.backendTLS |
Send TLS to the backend. |
mcp.policies.backendTLS.cert |
|
mcp.policies.backendTLS.key |
|
mcp.policies.backendTLS.root |
|
mcp.policies.backendTLS.hostname |
|
mcp.policies.backendTLS.insecure |
|
mcp.policies.backendTLS.insecureHost |
|
mcp.policies.backendTLS.alpn |
|
mcp.policies.backendTLS.subjectAltNames |
|
mcp.policies.backendAuth |
Authenticate to the backend. |
mcp.policies.backendAuth.(any)(1)passthrough |
|
mcp.policies.backendAuth.(any)(1)key |
|
mcp.policies.backendAuth.(any)(1)key.(any)file |
|
mcp.policies.backendAuth.(any)(1)gcp |
|
mcp.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.backendAuth.(any)(1)gcp.(any)audience |
Audience for the token. If not set, the destination host will be used. |
mcp.policies.backendAuth.(any)(1)gcp.(any)type |
|
mcp.policies.backendAuth.(any)(1)aws |
|
mcp.policies.backendAuth.(any)(1)aws.(any)accessKeyId |
|
mcp.policies.backendAuth.(any)(1)aws.(any)secretAccessKey |
|
mcp.policies.backendAuth.(any)(1)aws.(any)region |
|
mcp.policies.backendAuth.(any)(1)aws.(any)sessionToken |
|
mcp.policies.backendAuth.(any)(1)azure |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.tenant_id |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_id |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)clientSecret.client_secret |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)clientId |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)objectId |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)managedIdentity.userAssignedIdentity.(any)(1)resourceId |
|
mcp.policies.backendAuth.(any)(1)azure.(1)explicitConfig.(1)workloadIdentity |
|
mcp.policies.backendAuth.(any)(1)azure.(1)developerImplicit |
|
mcp.policies.localRateLimit |
Rate limit incoming requests. State is kept local. |
mcp.policies.localRateLimit[].maxTokens |
|
mcp.policies.localRateLimit[].tokensPerFill |
|
mcp.policies.localRateLimit[].fillInterval |
|
mcp.policies.localRateLimit[].type |
|
mcp.policies.remoteRateLimit |
Rate limit incoming requests. State is managed by a remote server. |
mcp.policies.remoteRateLimit.(any)(1)service |
|
mcp.policies.remoteRateLimit.(any)(1)service.name |
|
mcp.policies.remoteRateLimit.(any)(1)service.name.namespace |
|
mcp.policies.remoteRateLimit.(any)(1)service.name.hostname |
|
mcp.policies.remoteRateLimit.(any)(1)service.port |
|
mcp.policies.remoteRateLimit.(any)(1)host |
Hostname or IP address |
mcp.policies.remoteRateLimit.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
mcp.policies.remoteRateLimit.(any)domain |
|
mcp.policies.remoteRateLimit.(any)descriptors |
|
mcp.policies.remoteRateLimit.(any)descriptors[].entries |
|
mcp.policies.remoteRateLimit.(any)descriptors[].entries[].key |
|
mcp.policies.remoteRateLimit.(any)descriptors[].entries[].value |
|
mcp.policies.remoteRateLimit.(any)descriptors[].type |
|
mcp.policies.remoteRateLimit.(any)timeout |
Timeout for the request |
mcp.policies.jwtAuth |
Authenticate incoming JWT requests. |
mcp.policies.jwtAuth.(any)(any)mode |
|
mcp.policies.jwtAuth.(any)(any)providers |
|
mcp.policies.jwtAuth.(any)(any)providers[].issuer |
|
mcp.policies.jwtAuth.(any)(any)providers[].audiences |
|
mcp.policies.jwtAuth.(any)(any)providers[].jwks |
|
mcp.policies.jwtAuth.(any)(any)providers[].jwks.(any)file |
|
mcp.policies.jwtAuth.(any)(any)providers[].jwks.(any)url |
|
mcp.policies.jwtAuth.(any)(any)mode |
|
mcp.policies.jwtAuth.(any)(any)issuer |
|
mcp.policies.jwtAuth.(any)(any)audiences |
|
mcp.policies.jwtAuth.(any)(any)jwks |
|
mcp.policies.jwtAuth.(any)(any)jwks.(any)file |
|
mcp.policies.jwtAuth.(any)(any)jwks.(any)url |
|
mcp.policies.basicAuth |
Authenticate incoming requests using Basic Authentication with htpasswd. |
mcp.policies.basicAuth.htpasswd |
.htpasswd file contents/reference |
mcp.policies.basicAuth.htpasswd.(any)file |
|
mcp.policies.basicAuth.realm |
Realm name for the WWW-Authenticate header |
mcp.policies.basicAuth.mode |
Validation mode for basic authentication |
mcp.policies.apiKey |
Authenticate incoming requests using API Keys |
mcp.policies.apiKey.keys |
List of API keys |
mcp.policies.apiKey.keys[].key |
|
mcp.policies.apiKey.keys[].metadata |
|
mcp.policies.apiKey.mode |
Validation mode for API keys |
mcp.policies.extAuthz |
Authenticate incoming requests by calling an external authorization server. |
mcp.policies.extAuthz.(any)(1)service |
|
mcp.policies.extAuthz.(any)(1)service.name |
|
mcp.policies.extAuthz.(any)(1)service.name.namespace |
|
mcp.policies.extAuthz.(any)(1)service.name.hostname |
|
mcp.policies.extAuthz.(any)(1)service.port |
|
mcp.policies.extAuthz.(any)(1)host |
Hostname or IP address |
mcp.policies.extAuthz.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
mcp.policies.extAuthz.(any)protocol |
The ext_authz protocol to use. Unless you need to integrate with an HTTP-only server, gRPC is recommended. |
mcp.policies.extAuthz.(any)protocol.(1)grpc |
|
mcp.policies.extAuthz.(any)protocol.(1)grpc.context |
Additional context to send to the authorization service. This maps to the context_extensions field of the request, and only allows static values. |
mcp.policies.extAuthz.(any)protocol.(1)grpc.metadata |
Additional metadata to send to the authorization service. This maps to the metadata_context.filter_metadata field of the request, and allows dynamic CEL expressions.If unset, by default the envoy.filters.http.jwt_authn key is set if the JWT policy is used as well, for compatibility. |
mcp.policies.extAuthz.(any)protocol.(1)http |
|
mcp.policies.extAuthz.(any)protocol.(1)http.path |
|
mcp.policies.extAuthz.(any)protocol.(1)http.redirect |
When using the HTTP protocol, and the server returns unauthorized, redirect to the URL resolved by the provided expression rather than directly returning the error. |
mcp.policies.extAuthz.(any)protocol.(1)http.includeResponseHeaders |
Specific headers from the authorization response will be copied into the request to the backend. |
mcp.policies.extAuthz.(any)protocol.(1)http.addRequestHeaders |
Specific headers to add in the authorization request (empty = all headers), based on the expression |
mcp.policies.extAuthz.(any)protocol.(1)http.metadata |
Metadata to include under the extauthz variable, based on the authorization response. |
mcp.policies.extAuthz.(any)failureMode |
Behavior when the authorization service is unavailable or returns an error |
mcp.policies.extAuthz.(any)failureMode.(1)denyWithStatus |
|
mcp.policies.extAuthz.(any)includeRequestHeaders |
Specific headers to include in the authorization request. If unset, the gRPC protocol sends all request headers. The HTTP protocol sends only ‘Authorization’. |
mcp.policies.extAuthz.(any)includeRequestBody |
Options for including the request body in the authorization request |
mcp.policies.extAuthz.(any)includeRequestBody.maxRequestBytes |
Maximum size of request body to buffer (default: 8192) |
mcp.policies.extAuthz.(any)includeRequestBody.allowPartialMessage |
If true, send partial body when max_request_bytes is reached |
mcp.policies.extAuthz.(any)includeRequestBody.packAsBytes |
If true, pack body as raw bytes in gRPC |
mcp.policies.extAuthz.(any)timeout |
Timeout for the authorization request (default: 200ms) |
mcp.policies.extProc |
Extend agentgateway with an external processor |
mcp.policies.extProc.(any)(1)service |
|
mcp.policies.extProc.(any)(1)service.name |
|
mcp.policies.extProc.(any)(1)service.name.namespace |
|
mcp.policies.extProc.(any)(1)service.name.hostname |
|
mcp.policies.extProc.(any)(1)service.port |
|
mcp.policies.extProc.(any)(1)host |
Hostname or IP address |
mcp.policies.extProc.(any)(1)backend |
Explicit backend reference. Backend must be defined in the top level backends list |
mcp.policies.extProc.(any)failureMode |
Behavior when the ext_proc service is unavailable or returns an error |
mcp.policies.extProc.(any)metadataContext |
Additional metadata to send to the external processing service. Maps to the metadata_context.filter_metadata field in ProcessingRequest, and allows dynamic CEL expressions. |
mcp.policies.extProc.(any)requestAttributes |
Maps to the request attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
mcp.policies.extProc.(any)responseAttributes |
Maps to the response attributes field in ProcessingRequest, and allows dynamic CEL expressions. |
mcp.policies.transformations |
Modify requests and responses |
mcp.policies.transformations.request |
|
mcp.policies.transformations.request.add |
|
mcp.policies.transformations.request.set |
|
mcp.policies.transformations.request.remove |
|
mcp.policies.transformations.request.body |
|
mcp.policies.transformations.response |
|
mcp.policies.transformations.response.add |
|
mcp.policies.transformations.response.set |
|
mcp.policies.transformations.response.remove |
|
mcp.policies.transformations.response.body |
|
mcp.policies.csrf |
Handle CSRF protection by validating request origins against configured allowed origins. |
mcp.policies.csrf.additionalOrigins |
|
mcp.policies.timeout |
Timeout requests that exceed the configured duration. |
mcp.policies.timeout.requestTimeout |
|
mcp.policies.timeout.backendRequestTimeout |
|
mcp.policies.retry |
Retry matching requests. |
mcp.policies.retry.attempts |
|
mcp.policies.retry.backoff |
|
mcp.policies.retry.codes |
