Pomerium

# Configuration Settings

Pomerium can be configured using a configuration file (YAML (opens new window)/JSON (opens new window)/TOML (opens new window)) or environmental variables (opens new window). In general, environmental variable keys are identical to config file keys but are uppercase. If you are coming from a kubernetes or docker background this should feel familiar. If not, check out the following primers.

Using both environmental variables (opens new window) and config file keys is allowed and encouraged (for instance, secret keys are probably best set as environmental variables). However, if duplicate configuration keys are found, environment variables take precedence.

TIP

Pomerium can hot-reload route configuration details, authorization policy, certificates, and other proxy settings.

# Shared Settings

These configuration variables are shared by all services, in all service modes.

# Address

  • Environmental Variable: ADDRESS
  • Config File Key: address
  • Type: string
  • Example: :443, :8443
  • Default: :443
  • Required

Address specifies the host and port to serve HTTP requests from. If empty, :443 is used. Note, in all-in-one deployments, gRPC traffic will be served on loopback on port :5443.

# Autocert

  • Environmental Variable: AUTOCERT
  • Config File Key: autocert
  • Type: bool
  • Optional

Turning on autocert allows Pomerium to automatically retrieve, manage, and renew public facing TLS certificates from Let's Encrypt (opens new window) which includes managed routes and the authenticate service. Autocert Directory must be used with Autocert must have a place to persist, and share certificate data between services. Note that autocert also provides OCSP stapling (opens new window).

This setting can be useful in situations where you may not have Pomerium behind a TLS terminating ingress or proxy that is already handling your public certificates on your behalf.

WARNING

By using autocert, you agree to the Let's Encrypt Subscriber Agreement (opens new window). There are strict usage limits (opens new window) per domain you should be aware of. Consider testing with autocert_use_staging first.

WARNING

Autocert requires that ports 80/443 be accessible from the internet in order to complete a TLS-ALPN-01 challenge (opens new window).

# Autocert CA

  • Environmental Variable: AUTOCERT_CA
  • Config File Key: autocert_ca
  • Type: string containing the directory URL of an ACME CA (e.g. https://acme.zerossl.com/v2/DV90 for ZeroSSL)
  • Optional

Autocert CA is the directory URL of the ACME CA to use when requesting certificates.

TIP

This will overrule the "Autocert Use Staging" setting if set.

# Autocert Email

  • Environmental Variable: AUTOCERT_EMAIL
  • Config File Key: autocert_email
  • Type: string containing the email address to use when registering an account
  • Optional

Autocert Email is the email address to use when requesting certificates from an ACME CA.

TIP

The CA may contact you at this address, for example when a certificate expires.

# Autocert Must-Staple

  • Environmental Variable: AUTOCERT_MUST_STAPLE
  • Config File Key: autocert_must_staple
  • Type: bool
  • Optional

If true, force autocert to request a certificate with the status_request extension (commonly called Must-Staple). This allows the TLS client (id est the browser) to fail immediately if the TLS handshake doesn't include OCSP stapling information. This setting is only used when Autocert is true.

TIP

This setting will only take effect when you request or renew your certificates.

For more details, please see RFC7633 (opens new window) .

# Autocert Directory

  • Environmental Variable: either AUTOCERT_DIR

  • Config File Key: autocert_dir

  • Type: string pointing to the path of the directory

  • Required if using Autocert setting

  • Default:

Autocert directory is the path which autocert will store x509 certificate data.

# Autocert Use Staging

  • Environmental Variable: AUTOCERT_USE_STAGING
  • Config File Key: autocert_use_staging
  • Type: bool
  • Optional

Let's Encrypt has strict usage limits (opens new window). Enabling this setting allows you to use Let's Encrypt's staging environment (opens new window) which has much more lax usage limits.

# Autocert EAB Key ID

  • Environmental Variable: AUTOCERT_EAB_KEY_ID
  • Config File Key: autocert_eab_key_id
  • Type: string containing the identifier for an ACME EAB key to use
  • Optional

Autocert EAB Key ID is the key identifier when requesting a certificate from a CA with External Account Binding enabled.

For more information, please see RFC8555-#7.3.4 (opens new window).

# Autocert EAB MAC Key

  • Environmental Variable: AUTOCERT_EAB_MAC_KEY
  • Config File Key: autocert_eab_mac_key
  • Type: string containing a base64url-encoded secret key
  • Optional

Autocert EAB MAC Key is the base64url-encoded secret key corresponding to the Autocert EAB Key ID.

This setting is required when Autocert EAB Key ID is set.

# Autocert Trusted Certificate Authority

  • Environment Variable: AUTOCERT_TRUSTED_CA / AUTOCERT_TRUSTED_CA_FILE
  • Config File Key: autocert_trusted_ca / autocert_trusted_ca_file
  • Type: base64 encoded (opens new window) string or relative file location
  • Optional

The Autocert Trusted Certificate Authority is the x509 CA (bundle) used when communicating with a CA supporting the ACME protocol. If not set, the system trusted roots will be used to verify TLS connections to the ACME CA.

# Certificates

  • Config File Key: certificates (not yet settable using environmental variables)
  • Config File Key: certificate / certificate_key
  • Config File Key: certificate_file / certificate_key_file
  • Environmental Variable: CERTIFICATE / CERTIFICATE_KEY
  • Environmental Variable: CERTIFICATE_FILE / CERTIFICATE_KEY_FILE
  • Type: array of relative file locations string
  • Type: base64 encoded (opens new window) string
  • Type: certificate relative file location string
  • Required (if insecure not set)

Certificates are the x509 public-key and private-key used to establish secure HTTP and gRPC connections. Any combination of the above can be used together, and are additive. You can also use any of these settings in conjunction with Autocert to get OCSP stapling.

Certificates loaded into Pomerium from these config values are used to attempt secure connections between end users and services, between Pomerium services, and to upstream endpoints.

For example, if specifying multiple certificates at once:

certificates:
  - cert: "$HOME/.acme.sh/authenticate.example.com_ecc/fullchain.cer"
    key: "$HOME/.acme.sh/authenticate.example.com_ecc/authenticate.example.com.key"
  - cert: "$HOME/.acme.sh/verify.example.com_ecc/fullchain.cer"
    key: "$HOME/.acme.sh/verify.example.com_ecc/verify.example.com.key"
  - cert: "$HOME/.acme.sh/prometheus.example.com_ecc/fullchain.cer"
    key: "$HOME/.acme.sh/prometheus.example.com_ecc/prometheus.example.com.key"

Or to set a single certificate and key covering multiple domains and/or a wildcard subdomain:

certificate_file: "$HOME/.acme.sh/*.example.com/fullchain.crt"
certificate_key:  "$HOME/.acme.sh/*.example.com/*.example.com.key"

Note: Pomerium will check your system's trust/key store for valid certificates first. If your certificate solution imports into the system store, you don't need to also specify them with these configuration keys.

# Client Certificate Authority

  • Environment Variable: CLIENT_CA / CLIENT_CA_FILE
  • Config File Key: client_ca / client_ca_file
  • Type: base64 encoded (opens new window) string or relative file location
  • Optional

The Client Certificate Authority is the x509 public-key used to validate mTLS (opens new window) client certificates. If not set, no client certificate will be required.

# Client CRL

  • Environment Variable: CLIENT_CRL / CLIENT_CRL_FILE
  • Config File Key: client_crl / client_crl_file
  • Type: base64 encoded (opens new window) string or relative file location
  • Optional

The Client CRL is the certificate revocation list (opens new window) (in PEM format) for client certificates. If not set, no CRL will be used.

  • Environmental Variable: COOKIE_NAME
  • Config File Key: cookie_name
  • Type: string
  • Default: _pomerium

The name of the session cookie sent to clients.

Secret used to encrypt and sign session cookies. You can generate a random key with head -c32 /dev/urandom | base64.

  • Environmental Variable: COOKIE_DOMAIN
  • Config File Key: cookie_domain
  • Type: string
  • Example: localhost.pomerium.io
  • Optional

The scope of session cookies issued by Pomerium.

# HTTPS only

  • Environmental Variable: COOKIE_SECURE
  • Config File Key: cookie_secure
  • Type: bool
  • Default: true

If true, instructs browsers to only send user session cookies over HTTPS.

WARNING

Setting this to false may result in session cookies being sent in cleartext.

# Javascript Security

  • Environmental Variable: COOKIE_HTTP_ONLY
  • Config File Key: cookie_http_only
  • Type: bool
  • Default: true

If true, prevents javascript in browsers from reading user session cookies.

WARNING

Setting this to false enables hostile javascript to steal session cookies and impersonate users.

# Expiration

Sets the lifetime of session cookies. After this interval, users must reauthenticate.

# Debug

  • Environmental Variable: POMERIUM_DEBUG
  • Config File Key: pomerium_debug
  • Type: bool
  • Default: false

WARNING

Enabling the debug flag could result in sensitive information being logged!!!

By default, JSON encoded logs are produced. Debug enables colored, human-readable logs to be streamed to standard out (opens new window). In production, it is recommended to be set to false.

For example, if true

10:37AM INF cmd/pomerium version=v0.0.1-dirty+ede4124
10:37AM INF proxy: new route from=verify.localhost.pomerium.io to=https://verify.pomerium.com
10:37AM INF proxy: new route from=ssl.localhost.pomerium.io to=http://neverssl.com
10:37AM INF proxy/authenticator: grpc connection OverrideCertificateName= addr=auth.localhost.pomerium.io:443

If false

{"level":"info","version":"v0.0.1-dirty+ede4124","time":"2019-02-18T10:41:03-08:00","message":"cmd/pomerium"}
{"level":"info","from":"verify.localhost.pomerium.io","to":"https://verify.pomerium.com","time":"2019-02-18T10:41:03-08:00","message":"proxy: new route"}
{"level":"info","from":"ssl.localhost.pomerium.io","to":"http://neverssl.com","time":"2019-02-18T10:41:03-08:00","message":"proxy: new route"}
{"level":"info","OverrideCertificateName":"","addr":"auth.localhost.pomerium.io:443","time":"2019-02-18T10:41:03-08:00","message":"proxy/authenticator: grpc connection"}

# Forward Auth

  • Environmental Variable: FORWARD_AUTH_URL
  • Config File Key: forward_auth_url
  • Type: URL (must contain a scheme and hostname)
  • Example: https://forwardauth.corp.example.com
  • Resulting Verification URL: https://forwardauth.corp.example.com/?uri={URL-TO-VERIFY}
  • Optional

Forward authentication creates an endpoint that can be used with third-party proxies that do not have rich access control capabilities (nginx (opens new window), nginx-ingress (opens new window), ambassador (opens new window), traefik (opens new window)). Forward authentication allows you to delegate authentication and authorization for each request to Pomerium.

# Request flow

pomerium forward auth request flow

# Examples

# NGINX Ingress

Some reverse-proxies, such as nginx split access control flow into two parts: verification and sign-in redirection. Notice the additional path /verify used for auth-url indicating to Pomerium that it should return a 401 instead of redirecting and starting the sign-in process.

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: verify
  annotations:
    kubernetes.io/ingress.class: "nginx"
    certmanager.k8s.io/issuer: "letsencrypt-prod"
    nginx.ingress.kubernetes.io/auth-url: https://forwardauth.corp.example.com/verify?uri=$scheme://$host$request_uri
    nginx.ingress.kubernetes.io/auth-signin: "https://forwardauth.corp.example.com/?uri=$scheme://$host$request_uri"
spec:
  tls:
    - hosts:
        - verify.corp.example.com
      secretName: quickstart-example-tls
  rules:
    - host: verify.corp.example.com
      http:
        paths:
          - path: /
            backend:
              serviceName: verify
              servicePort: 80

# Traefik docker-compose

If the forward_auth_url is also handled by Traefik, you will need to configure Traefik to trust the X-Forwarded-* headers as described in the documentation (opens new window).

version: "3"

services:
  traefik:
    # The official v2.2 Traefik docker image
    image: traefik:v2.2
    # Enables the web UI and tells Traefik to listen to docker
    command:
      - "--api.insecure=true"
      - "--providers.docker=true"
      - "--entrypoints.web.address=:80"
      - "--entrypoints.web.forwardedheaders.insecure=true"
    ports:
      # The HTTP port
      - "80:80"
      # The Web UI (enabled by --api.insecure=true)
      - "8080:8080"
    volumes:
      # So that Traefik can listen to the Docker events
      - /var/run/docker.sock:/var/run/docker.sock
  verify:
    # A container that exposes an API to show its IP address
    image: pomerium/verify:latest
    labels:
      - "traefik.http.routers.verify.rule=Host(`verify.corp.example.com`)"
      # Create a middleware named `foo-add-prefix`
      - "traefik.http.middlewares.test-auth.forwardauth.authResponseHeaders=x-pomerium-claim-email,x-pomerium-claim-id,x-pomerium-claim-groups,x-pomerium-jwt-assertion"
      - "traefik.http.middlewares.test-auth.forwardauth.address=http://forwardauth.corp.example.com/?uri=https://verify.corp.example.com"
      - "traefik.http.routers.verify.middlewares=test-auth@docker"

# Global Timeouts

  • Environmental Variables: TIMEOUT_READ TIMEOUT_WRITE TIMEOUT_IDLE
  • Config File Key: timeout_read timeout_write timeout_idle
  • Type: Go Duration (opens new window) string
  • Example: TIMEOUT_READ=30s
  • Defaults: TIMEOUT_READ=30s TIMEOUT_WRITE=0 TIMEOUT_IDLE=5m

Timeouts set the global server timeouts. Timeouts can also be set for individual routes.

cloudflare blog on timeouts

For a deep dive on timeout values see these (opens new window) two (opens new window) excellent blog posts.

# GRPC Options

# GRPC Address

  • Environmental Variable: GRPC_ADDRESS
  • Config File Key: grpc_address
  • Type: string
  • Example: :443, :8443
  • Default: :443 or :5443 if in all-in-one mode

gRPC Address specifies the host and port to serve gRPC requests from.

# GRPC Insecure

  • Environmental Variable: GRPC_INSECURE
  • Config File Key: grpc_insecure
  • Type: bool

This setting disables transport security for gRPC communication. If running in all-in-one mode, defaults to true as communication will run over localhost's own socket.

# GRPC Client Timeout

Maximum time before canceling an upstream gRPC request. During transient failures, the proxy will retry upstreams for this duration. You should leave this high enough to handle backend service restart and rediscovery so that client requests do not fail.

# GRPC Client DNS RoundRobin

  • Environmental Variable: GRPC_CLIENT_DNS_ROUNDROBIN
  • Config File Key: grpc_client_dns_roundrobin
  • Type: bool
  • Default: true

Enable gRPC DNS based round robin load balancing. This method uses DNS to resolve endpoints and does client side load balancing of all addresses returned by the DNS record. Do not disable unless you have a specific use case.

# HTTP Redirect Address

  • Environmental Variable: HTTP_REDIRECT_ADDR
  • Config File Key: http_redirect_addr
  • Type: string
  • Example: :80, :8080
  • Optional

If set, the HTTP Redirect Address specifies the host and port to redirect http to https traffic on. If unset, no redirect server is started.

# Insecure Server

  • Environmental Variable: INSECURE_SERVER
  • Config File Key: insecure_server
  • Type: bool
  • Required if certificates unset

Turning on insecure server mode will result in pomerium starting, and operating without any protocol encryption in transit.

This setting can be useful in a situation where you have Pomerium behind a TLS terminating ingress or proxy. However, even in that case, it is highly recommended to use TLS to protect the confidentiality and integrity of service communication even behind the ingress using self-signed certificates or an internal CA. Please see our helm-chart for an example of just that.

WARNING

Pomerium should never be exposed to the internet without TLS encryption.

# DNS Lookup Family

  • Environmental Variable: DNS_LOOKUP_FAMILY
  • Config File Key: dns_lookup_family
  • Type: string
  • Options: V4_ONLY V6_ONLY AUTO
  • Optional

The DNS IP address resolution policy. If not specified, the value defaults to AUTO.

# Log Level

  • Environmental Variable: LOG_LEVEL
  • Config File Key: log_level
  • Type: string
  • Options: debug info warn error
  • Default: debug

Log level sets the global logging level for pomerium. Only logs of the desired level and above will be logged.

# Metrics Address

  • Environmental Variable: METRICS_ADDRESS
  • Config File Key: metrics_address
  • Type: string
  • Example: :9090, 127.0.0.1:9090
  • Default: disabled
  • Optional

Expose a prometheus endpoint on the specified port.

WARNING

Use with caution: the endpoint can expose frontend and backend server names or addresses. Do not externally expose the metrics if this is sensitive information.

# Pomerium Metrics Tracked

Each metric exposed by Pomerium has a pomerium prefix, which is omitted in the table below for brevity.

Name Type Description
build_info Gauge Pomerium build metadata by git revision, service, version and goversion
config_checksum_int64 Gauge Currently loaded configuration checksum by service
config_last_reload_success Gauge Whether the last configuration reload succeeded by service
config_last_reload_success_timestamp Gauge The timestamp of the last successful configuration reload by service
grpc_client_request_duration_ms Histogram GRPC client request duration by service
grpc_client_request_size_bytes Histogram GRPC client request size by service
grpc_client_requests_total Counter Total GRPC client requests made by service
grpc_client_response_size_bytes Histogram GRPC client response size by service
grpc_server_request_duration_ms Histogram GRPC server request duration by service
grpc_server_request_size_bytes Histogram GRPC server request size by service
grpc_server_requests_total Counter Total GRPC server requests made by service
grpc_server_response_size_bytes Histogram GRPC server response size by service
http_client_request_duration_ms Histogram HTTP client request duration by service
http_client_request_size_bytes Histogram HTTP client request size by service
http_client_requests_total Counter Total HTTP client requests made by service
http_client_response_size_bytes Histogram HTTP client response size by service
http_server_request_duration_ms Histogram HTTP server request duration by service
http_server_request_size_bytes Histogram HTTP server request size by service
http_server_requests_total Counter Total HTTP server requests handled by service
http_server_response_size_bytes Histogram HTTP server response size by service
redis_conns Gauge Number of total connections in the pool
redis_idle_conns Gauge Total number of times free connection was found in the pool
redis_wait_count_total Counter Total number of connections waited for
redis_wait_duration_ms_total Counter Total time spent waiting for connections
storage_operation_duration_ms Histogram Storage operation duration by operation, result, backend and service

# Identity Manager

Identity manager metrics have pomerium_identity_manager prefix.

Name Type Description
last_refresh_timestamp Gauge Timestamp of last directory refresh operation.
session_refresh_error_timestamp Gauge Timestamp of last session refresh ended in an error.
session_refresh_errors Counter Session refresh error counter.
session_refresh_success Counter Session refresh success counter.
session_refresh_success_timestamp Gauge Timestamp of last successful session refresh.
user_group_refresh_error_timestamp Gauge Timestamp of last user group refresh ended in an error.
user_group_refresh_errors Counter User group refresh error counter.
user_group_refresh_success Counter User group refresh success counter.
user_group_refresh_success_timestamp Gauge Timestamp of last group successful user refresh.
user_refresh_error_timestamp Gauge Timestamp of last user refresh ended in an error.
user_refresh_errors Counter User refresh error counter.
user_refresh_success Counter User refresh success counter.
user_refresh_success_timestamp Gauge Timestamp of last successful user refresh.

# Envoy Proxy Metrics

As of v0.9, Pomerium uses envoy (opens new window) for the data plane. As such, proxy related metrics are sourced from envoy, and use envoy's internal stats data model (opens new window). Please see Envoy's documentation for information about specific metrics.

All metrics coming from envoy will be labeled with service="pomerium" or service="pomerium-proxy", depending if you're running all-in-one or distributed service mode and have pomerium prefix added to the standard envoy metric name.

# Metrics Basic Authentication

  • Environmental Variable: METRICS_BASIC_AUTH
  • Config File Key: metrics_basic_auth
  • Type: base64 encoded string of username:password
  • Example: eDp5 (for username: x, and password: y)
  • Default: ``
  • Optional

Require Basic HTTP Authentication (opens new window) to access the metrics endpoint.

To support this in Prometheus, consult the basic_auth option in the scrape_config (opens new window) documentation.

# Metrics Certificate

  • Config File Key: metrics_certificate / metrics_certificate_key
  • Config File Key: metrics_certificate_file / metrics_certificate_key_file
  • Environmental Variable: METRICS_CERTIFICATE / METRICS_CERTIFICATE_KEY
  • Environmental Variable: METRICS_CERTIFICATE_FILE / METRICS_CERTIFICATE_KEY_FILE
  • Type: base64 encoded (opens new window) string
  • Type: certificate relative file location string
  • Optional

Certificates are the x509 public-key and private-key used to secure the metrics endpoint.

# Metrics Client Certificate Authority

  • Environment Variable: METRICS_CLIENT_CA / METRICS_CLIENT_CA_FILE
  • Config File Key: metrics_client_ca / metrics_client_ca_file
  • Type: base64 encoded (opens new window) string or relative file location
  • Optional

The Client Certificate Authority is the x509 public-key used to validate mTLS (opens new window) client certificates for the metrics endpoint. If not set, no client certificate will be required.

# Proxy Log Level

  • Environmental Variable: PROXY_LOG_LEVEL
  • Config File Key: proxy_log_level
  • Type: string
  • Options: debug info warn error
  • Default: value of log_level or debug if both are unset

Proxy log level sets the logging level for the Pomerium Proxy service access logs. Only logs of the desired level and above will be logged.

# Service Mode

  • Environmental Variable: SERVICES
  • Config File Key: services
  • Type: string
  • Default: all
  • Options: all authenticate authorize databroker or proxy

Service mode sets which service(s) to run. If testing, you may want to set to all and run pomerium in "all-in-one mode." In production, you'll likely want to spin up several instances of each service mode for high availability.

# Shared Secret

Shared Secret is the base64 encoded 256-bit key used to mutually authenticate requests between services. It's critical that secret keys are random, and stored safely. Use a key management system or /dev/urandom to generate a key. For example:

head -c32 /dev/urandom | base64

# Tracing

Tracing tracks the progression of a single user request as it is handled by Pomerium.

Each unit of work is called a Span in a trace. Spans include metadata about the work, including the time spent in the step (latency), status, time events, attributes, links. You can use tracing to debug errors and latency issues in your applications, including in downstream connections.

# Shared Tracing Settings

Config Key Description Required
tracing_provider The name of the tracing provider. (e.g. jaeger, zipkin) ✅
tracing_sample_rate Percentage of requests to sample in decimal notation. Default is 0.0001, or .01% ❌

# Datadog

Datadog is a real-time monitoring system that supports distributed tracing and monitoring.

Config Key Description Required
tracing_datadog_address host:port address of the Datadog Trace Agent. Defaults to localhost:8126 ❌

# Jaeger (partial)

Warning At this time, Jaeger protocol does not capture spans inside the Proxy service. Please use Zipkin protocol with Jaeger for full support.

Jaeger (opens new window) is a distributed tracing system released as open source by Uber Technologies. It is used for monitoring and troubleshooting microservices-based distributed systems, including:

  • Distributed context propagation
  • Distributed transaction monitoring
  • Root cause analysis
  • Service dependency analysis
  • Performance / latency optimization
Config Key Description Required
tracing_jaeger_collector_endpoint Url to the Jaeger HTTP Thrift collector. ✅
tracing_jaeger_agent_endpoint Send spans to jaeger-agent at this address. ✅

# Zipkin

Zipkin is an open source distributed tracing system and protocol.

Many tracing backends support zipkin either directly or through intermediary agents, including Jaeger. For full tracing support, we recommend using the Zipkin tracing protocol.

Config Key Description Required
tracing_zipkin_endpoint Url to the Zipkin HTTP endpoint. ✅

# Example

jaeger example trace

# Use Proxy Protocol

  • Environment Variable: USE_PROXY_PROTOCOL
  • Config File Key: use_proxy_protocol
  • Type: bool
  • Optional

Setting use_proxy_protocol will configure Pomerium to require the HAProxy proxy protocol (opens new window) on incoming connections. Versions 1 and 2 of the protocol are supported.

# Envoy Bootstrap Options

  • Environment Variable: ENVOY_ADMIN_ADDRESS, ENVOY_ADMIN_ACCESS_LOG_PATH, ENVOY_ADMIN_PROFILE_PATH, ENVOY_BIND_CONFIG_FREEBIND, ENVOY_BIND_CONFIG_SOURCE_ADDRESS
  • Config File Keys: envoy_admin_address, envoy_admin_access_log_path, envoy_admin_profile_path, envoy_bind_config_freebind, envoy_bind_config_source_address
  • Type: string
  • Optional

The envoy_admin keys customize Envoy's bootstrap configuration (opens new window). The envoy_bind_config keys modify the ClusterManager (opens new window) configuration. These options cannot be modified at runtime.

# Authenticate Service

# Authenticate Callback Path

  • Environmental Variable: AUTHENTICATE_CALLBACK_PATH
  • Config File Key: authenticate_callback_path
  • Type: string
  • Default: /oauth2/callback
  • Optional

Authenticate callback path sets the path at which the authenticate service receives callback responses from your identity provider. The value must exactly match one of the authorized redirect URIs for the OAuth 2.0 client.

This value is referred to as the redirect_url in the OpenIDConnect (opens new window) and OAuth2 specs.

See also:

# Authenticate Service URL

  • Environmental Variable: AUTHENTICATE_SERVICE_URL
  • Config File Key: authenticate_service_url
  • Type: URL
  • Required
  • Example: https://authenticate.corp.example.com

Authenticate Service URL is the externally accessible URL for the authenticate service.

# Authenticate Internal Service URL

  • Environmental Variable: AUTHENTICATE_INTERNAL_SERVICE_URL
  • Config File Key: authenticate_internal_service_url
  • Type: URL
  • Required
  • Example: https://authenticate.internal

Authenticate Internal Service URL overrides authenticate_service_url when determining the TLS certificate and hostname for the authenticate service to listen with.

# Identity Provider Client ID

  • Environmental Variable: IDP_CLIENT_ID
  • Config File Key: idp_client_id
  • Type: string
  • Required

Client ID is the OAuth 2.0 Client Identifier retrieved from your identity provider. See your identity provider's documentation, and our identity provider docs for details.

# Identity Provider Client Secret

  • Environmental Variable: IDP_CLIENT_SECRET
  • Config File Key: idp_client_secret
  • Type: string
  • Required

Client Secret is the OAuth 2.0 Secret Identifier retrieved from your identity provider. See your identity provider's documentation, and our identity provider docs for details.

# Identity Provider Name

  • Environmental Variable: IDP_PROVIDER
  • Config File Key: idp_provider
  • Type: string
  • Required
  • Options: auth0 azure google okta onelogin or oidc

Provider is the short-hand name of a built-in OpenID Connect (oidc) identity provider to be used for authentication. To use a generic provider,set to oidc.

See identity provider for details.

# Identity Provider Scopes

  • Environmental Variable: IDP_SCOPES
  • Config File Key: idp_scopes
  • Type: list of string
  • Default: oidc,profile, email, offline_access (typically)
  • Optional for built-in identity providers.

Identity provider scopes correspond to access privilege scopes as defined in Section 3.3 of OAuth 2.0 RFC6749. The scopes associated with Access Tokens determine what resources will be available when they are used to access OAuth 2.0 protected endpoints.

WARNING

If you are using a built-in provider, you probably don't want to set customized scopes.

WARNING

Some providers, like Amazon Cognito, do not support the offline_access scope.

# Identity Provider Service Account

  • Environmental Variable: IDP_SERVICE_ACCOUNT
  • Config File Key: idp_service_account
  • Type: string
  • Required for group based policies (most configurations)

The identity provider service account setting is used to query associated identity information from your identity provider. This is a provider specific value and is not required for all providers. For example, when using Okta this value will be an Okta API key, and for an OIDC provider that provides groups as a claim, this value will be empty.

WARNING

If you plan to write authorization policies using groups, or any other data that exists in your identity provider's directory service, this setting is mandatory.

# Identity Provider URL

  • Environmental Variable: IDP_PROVIDER_URL
  • Config File Key: idp_provider_url
  • Type: string
  • Required, depending on provider

Provider URL is the base path to an identity provider's OpenID connect discovery document (opens new window). For example, google's URL would be https://accounts.google.com for their discover document (opens new window).

# Identity Provider Request Params

  • Environmental Variable: IDP_REQUEST_PARAMS
  • Config File Key: idp_request_params
  • Type: map of strings key value pairs
  • Optional

Request parameters to be added as part of a signin request using OAuth2 code flow.

For more information see:

# Identity Provider Refresh Directory Settings

  • Environmental Variables: IDP_REFRESH_DIRECTORY_INTERVAL IDP_REFRESH_DIRECTORY_TIMEOUT
  • Config File Key: idp_refresh_directory_interval idp_refresh_directory_timeout
  • Type: Go Duration (opens new window) string
  • Example: IDP_REFRESH_DIRECTORY_INTERVAL=30m
  • Defaults: IDP_REFRESH_DIRECTORY_INTERVAL=10m IDP_REFRESH_DIRECTORY_TIMEOUT=1m

Refresh directory interval is the time that pomerium will sync your IDP diretory, while refresh directory timeout is the maximum time allowed each run.

WARNING

Use it at your own risk, if you set a too low value, you may reach IDP API rate limit.

# Proxy Service

# Authenticate Service URL

  • Environmental Variable: AUTHENTICATE_SERVICE_URL
  • Config File Key: authenticate_service_url
  • Type: URL
  • Required
  • Example: https://authenticate.corp.example.com

Authenticate Service URL is the externally accessible URL for the authenticate service.

# Authorize Service URL

  • Environmental Variable: AUTHORIZE_SERVICE_URL
  • Config File Key: authorize_service_url
  • Type: URL
  • Required; inferred in all-in-one mode to be localhost.
  • Example: https://pomerium-authorize-service.default.svc.cluster.local or https://localhost:5443

Authorize Service URL is the location of the internally accessible Authorize service. NOTE: Unlike authenticate, authorize has no publicly accessible http handlers so this setting is purely for gRPC communication.

If your load balancer does not support gRPC pass-through you'll need to set this value to an internally routable location (https://pomerium-authorize-service.default.svc.cluster.local) instead of an externally routable one (https://authorize.corp.example.com).

# Authorize Internal Service URL

  • Environmental Variable: AUTHORIZE_INTERNAL_SERVICE_URL
  • Config File Key: authorize_internal_service_url
  • Type: URL
  • Required; inferred in all-in-one mode to be localhost.
  • Example: https://pomerium-authorize-service.default.svc.cluster.local or https://localhost:5443

Authorize Internal Service URL overrides authorize_service_url when determining the TLS certificate for the authorize service to listen with.

# Certificate Authority

  • Environmental Variable: CERTIFICATE_AUTHORITY or CERTIFICATE_AUTHORITY_FILE
  • Config File Key: certificate_authority or certificate_authority_file
  • Type: base64 encoded (opens new window) string or relative file location
  • Optional

Certificate Authority is set when behind-the-ingress service communication uses custom or self-signed certificates.

WARNING

Be sure to include the intermediary certificate.

# Default Upstream Timeout

  • Environmental Variable: DEFAULT_UPSTREAM_TIMEOUT
  • Config File Key: default_upstream_timeout
  • Type: Duration (opens new window) string
  • Example: 10m, 1h45m
  • Default: 30s

Default Upstream Timeout is the default timeout applied to a proxied route when no timeout key is specified by the policy.

# Set Response Headers

  • Environmental Variable: SET_RESPONSE_HEADERS

  • Config File Key: set_response_headers

  • Type: map of strings key value pairs

  • Examples:

    • Comma Separated: X-Content-Type-Options:nosniff,X-Frame-Options:SAMEORIGIN

    • JSON: '{"X-Test": "X-Value"}'

    • YAML:

      set_response_headers:
  X-Test: X-Value

  • To disable: disable:true

  • Default :

    X-Content-Type-Options : nosniff,
X-Frame-Options:SAMEORIGIN,
X-XSS-Protection:1; mode=block,
Strict-Transport-Security:max-age=31536000; includeSubDomains; preload,

Set Response Headers specifies a mapping of HTTP Header (opens new window) to be added globally to all managed routes and pomerium's authenticate service.

By default, conservative secure HTTP headers (opens new window) are set:

  • max-age=31536000 instructs the browser to pin the certificate for a domain for a year. This helps prevent man-in-the-middle attacks, but can create issues when developing new environments with temporary certificates. See Troubleshooting - HSTS for more information.
  • includeSubDomains applies these rules to subdomains, which is how individual routes are defined.
  • preload instructs the browser to preload the certificate from an HSTS preload service if available. This means that the certificate can be loaded from an already-trusted secure connection, and the user never needs to connect to your domain without TLS.

pomerium security headers

See MDN Web Docs - Strict-Transport-Security (opens new window) for more information.

TIP

Several security-related headers are not set by default since doing so might break legacy sites. These include: Cross-Origin Resource Policy, Cross-Origin Opener Policy and Cross-Origin Embedder Policy. If possible users are encouraged to add these to set_response_headers or their downstream applications.

# JWT Claim Headers

  • Environmental Variable: JWT_CLAIMS_HEADERS
  • Config File Key: jwt_claims_headers
  • Type: slice of string
  • Example: email,groups, user
  • Optional

The JWT Claim Headers setting allows you to pass specific user session data down to upstream applications as HTTP request headers. Note, unlike the header x-pomerium-jwt-assertion these values are not signed by the authorization service.

Any claim in the pomerium session JWT can be placed into a corresponding header for upstream consumption. This claim information is sourced from your Identity Provider (IdP) and Pomerium's own session metadata. The header will have the following format:

X-Pomerium-Claim-{Name} where {Name} is the name of the claim requested.

This option also supports a nested object to customize the header name. For example:

jwt_claims_headers:
  X-Email: email

Will add an X-Email header with a value of the email claim.

Use this option if you previously relied on x-pomerium-authenticated-user-{email|user-id|groups}.

# Override Certificate Name

  • Environmental Variable: OVERRIDE_CERTIFICATE_NAME
  • Config File Key: override_certificate_name
  • Type: int
  • Optional
  • Example: *.corp.example.com if wild card or authenticate.corp.example.com/authorize.corp.example.com

Secure service communication can fail if the external certificate does not match the internally routed service hostname/SNI (opens new window). This setting allows you to override that value.

# Programmatic Redirect Domain Whitelist

  • Config File Key: programmatic_redirect_domain_whitelist
  • Type: array of string
  • Optional
  • Default: localhost

The programmatic redirect domain whitelist is used to restrict the allowed redirect URLs when using programmatic login. By default only localhost URLs are allowed.

# X-Forwarded-For HTTP Header

  • Environmental Variable: SKIP_XFF_APPEND
  • Config File Key: skip_xff_append
  • Type: bool
  • Default: false

Do not append proxy IP address to x-forwarded-for HTTP header. See Envoy (opens new window) docs for more detail.

# The number of trusted hops

  • Environmental Variable: XFF_NUM_TRUSTED_HOPS
  • Config File Key: xff_num_trusted_hops
  • Type: uint32
  • Default: 0

The number of trusted reverse proxies in front of pomerium. This affects x-forwarded-proto header and x-envoy-external-address header (opens new window), which reports tursted client address. Envoy (opens new window) docs for more detail.

# Codec Type

  • Environment Variable: CODEC_TYPE
  • Config File Key: codec_type
  • Type: string
  • Default: auto (http1 in all-in-one mode)

Specifies the codec to use for downstream connections. Either auto, http1 or http2.

When auto is specified the codec will be determined via TLS ALPN or protocol inference.

WARNING

With HTTP/2, browsers typically coalesce connections for the same IP address that use the same TLS certificate. For example, you may have authenticate.localhost.pomerium.io and example.localhost.pomerium.io using the same wildcard certificate (*.localhost.pomerium.io) and both pointing to 127.0.0.1. Your browser sees this and re-uses the initial connection it makes to example for authenticate. But unfortunately the routes necessary to handle authenticate don't exist on example so the proxy cannot handle the request.

If this happens Pomerium will respond with a 421 Misdirected Request status. Most browsers will attempt to make the request on a new HTTP/2 connection. However not all browsers implement this behavior (notably Safari), and users may end up seeing a blank page instead.

If you see this happen, there are several ways to mitigate the problem:

  1. Don't re-use TLS certificates for shared IP domains.
  2. Don't re-use IP addresses for shared TLS certificates.
  3. Don't use HTTP/2.

More details on this problem are available in Github Issue #2150 (opens new window).

# Data Broker Service

The databroker service is used for storing user session data.

# Data Broker Service URL

  • Environmental Variable: DATABROKER_SERVICE_URL or DATABROKER_SERVICE_URLS
  • Config File Key: databroker_service_url or databroker_service_urls
  • Type: URL
  • Example: https://databroker.corp.example.com
  • Default: in all-in-one mode, http://localhost:5443

The data broker service URL points to a data broker which is responsible for storing associated authorization context (e.g. sessions, users and user groups). Multiple URLs can be specified with databroker_service_urls.

By default, the databroker service uses an in-memory databroker.

To create your own data broker, implement the following gRPC interface:

For an example implementation, the in-memory database used by the databroker service can be found here:

# Data Broker Internal Service URL

  • Environmental Variable: DATABROKER_INTERNAL_SERVICE_URL or DATABROKER_INTERNAL_SERVICE_URLS
  • Config File Key: databroker_internal_service_url or databroker_internal_service_urls
  • Type: URL
  • Example: https://databroker.corp.example.com
  • Default: in all-in-one mode, http://localhost:5443

Data Broker Internal URL overrides databroker_service_url when determining the TLS certificate for the databroker service to listen with.

# Data Broker Storage Type

  • Environmental Variable: DATABROKER_STORAGE_TYPE
  • Config File Key: databroker_storage_type
  • Type: string
  • Optional
  • Example: redis,memory
  • Default: memory

The backend storage that databroker server will use.

# Data Broker Storage Connection String

  • Environmental Variable: DATABROKER_STORAGE_CONNECTION_STRING
  • Config File Key: databroker_storage_connection_string
  • Type: string
  • Required when storage type is redis
  • Example: "redis://localhost:6379/0", "rediss://localhost:6379/0"

The connection string that the databroker service will use to connect to storage backend.

For redis, the following URL types are supported:

  • simple: redis://[username:password@]host:port/[db]
  • sentinel: redis+sentinel://[:password@]host:port[,host2:port2,...]/[master_name[/db]][?param1=value1[&param2=value2&...]]
  • cluster: redis+cluster://[username:password@]host:port[,host2:port2,...]/[?param1=value1[&param2=value=2&...]]

You can also enable TLS with rediss://, rediss+sentinel:// and rediss+cluster://.

# Data Broker Storage Certificate File

  • Environment Variable: DATABROKER_STORAGE_CERT_FILE
  • Config File Key: databroker_storage_cert_file
  • Type: relative file location
  • Optional

The certificate used to connect to a storage backend.

# Data Broker Storage Certificate Key File

  • Environment Variable: DATABROKER_STORAGE_KEY_FILE
  • Config File Key: databroker_storage_key_file
  • Type: relative file location
  • Optional

The certificate key used to connect to a storage backend.

# Data Broker Storage Certificate Authority

  • Environment Variable: DATABROKER_STORAGE_CA_FILE
  • Config File Key: databroker_storage_ca_file
  • Type: relative file location
  • Optional

This setting defines the set of root certificates used when verifying storage server connections.

# Data Broker Storage TLS Skip Verify

  • Environment Variable: DATABROKER_STORAGE_TLS_SKIP_VERIFY
  • Config File Key: databroker_storage_tls_skip_verify
  • Type: relative file location
  • Optional

If set, the TLS connection to the storage backend will not be verified.

# Policy

  • Environmental Variable: POLICY
  • Config File Key: policy
  • Type: base64 encoded (opens new window) string or inline policy structure in config file
  • Deprecated: This key has been replaced with route.

WARNING

The policy field as a top-level configuration key has been replaced with routes. Moving forward, define policies within each defined route.

Existing policy definitions will currently behave as expected, but are deprecated and will be removed in a future version of Pomerium.

Policy contains route specific settings, and access control details. If you are configuring via POLICY environment variable, just the contents of the policy needs to be passed. If you are configuring via file, the policy should be present under the policy key. For example,

# This file contains only policy and route configuration details. Other
# configuration settings required by pomerium are excluded for clarity.
# See: https://www.pomerium.com/docs/reference/

#
# For a complete self contained configuration see : config.example.yaml.
# Or, mix and match a policy file (this) with env vars : config.example.env

# Proxied routes and per-route policies are defined in a policy block
# NOTA BENE: You must uncomment the below 'policy' key if you are loading policy as a file.
# policy:
- from: https://verify.localhost.pomerium.io
  to: http://localhost:8000
  allowed_domains:
    - pomerium.io
  cors_allow_preflight: true
  timeout: 30s
- from: https://external-verify.localhost.pomerium.io
  to: https://verify.pomerium.com
  allowed_domains:
    - gmail.com
- from: https://weirdlyssl.localhost.pomerium.io
  to: http://neverssl.com
  allowed_users:
    - user@example.com
  allowed_groups:
    - admins
    - developers
- from: https://hello.localhost.pomerium.io
  to: http://localhost:8080
  allowed_groups:
    - admins@pomerium.io

Policy routes are checked in the order they appear in the policy, so more specific routes should appear before less specific routes. For example:

policy:
  - from: http://from.example.com
    to: http://to.example.com
    prefix: /admin
    allowed_groups: ["superuser"]
  - from: http://from.example.com
    to: http://to.example.com
    allow_public_unauthenticated_access: true

In this example, an incoming request with a path prefix of /admin would be handled by the first route (which is restricted to superusers). All other requests for from.example.com would be handled by the second route (which is open to the public).

A list of configuration variables specific to policy follows Note that this also shares all configuration variables listed under routes, excluding policy and its child variables.

# Allowed Domains

  • yaml/json setting: allowed_domains
  • Type: list of string
  • Required
  • Example: pomerium.io , gmail.com

Allowed domains is a collection of whitelisted domains to authorize for a given route.

# Allowed Groups

  • yaml/json setting: allowed_groups
  • Type: list of string
  • Required
  • Example: admins , support@company.com

Allowed groups is a collection of whitelisted groups to authorize for a given route.

# Allowed IdP Claims

  • yaml/json setting: allowed_idp_claims
  • Type: map of strings lists
  • Required

Allowed IdP Claims is a collection of whitelisted claim key-value pairs to authorize for a given route.

This is useful if your identity provider has extra information about a user that is not in the directory. It can also be useful if you wish to use groups with the generic OIDC provider.

Example:

  - from: http://from.example.com
    to: http://to.example.com
    allowed_idp_claims:
      family_name:
        - Doe
        - Smith

This policy would match users with the family_name claim containing Smith or Doe.

Claims are represented as a map of strings to a list of values:

{
  "family_name": ["Doe"],
  "given_name": ["John"]
}
  • Nested maps are flattened: { "a": { "b": ["c"] } } becomes { "a.b": ["c"] }
  • Values are always a list: { "a": "b" } becomes { "a": ["b"] }

# Allowed Users

  • yaml/json setting: allowed_users
  • Type: list of string
  • Required
  • Example: alice@pomerium.io , bob@contractor.co

Allowed users is a collection of whitelisted users to authorize for a given route.

# Routes

  • Environment Variable: ROUTES
  • Config File Key: routes
  • Type: base64 encoded (opens new window) string or inline policy structure in config file
  • Required - While Pomerium will start without a route configured, it will not authorize or proxy any traffic until a route is defined. If configuring Pomerium for the Enterprise Console, define a route for the Console itself in Pomerium.

A route contains specific access and control definitions for a back-end service. Each route is a list item under the routes key.

Each route defines at minimum a from and to field, and a policy key defining authorization logic. Policies are defined using Pomerium Policy Language (PPL). Additional options are listed below.

# This file contains only route and policy configuration details. Other
# configuration settings required by pomerium are excluded for clarity.
# See: https://www.pomerium.io/docs/reference/

#
# For a complete self contained configuration see : config.example.yaml.
# Or, mix and match a policy file (this) with env vars : config.example.env

routes:
  - from: https://verify.localhost.pomerium.io
    to: http://localhost:8000
    policy:
      - allow:
          or:
            - domain:
                is: pomerium.io
    cors_allow_preflight: true
    timeout: 30s
  - from: https://external-verify.localhost.pomerium.io
    to: https://verify.pomerium.com
    policy:
      - allow:
          or:
            - domain:
                is: gmail.com
  - from: https://weirdlyssl.localhost.pomerium.io
    to: http://neverssl.com
    policy:
      - allow:
          or:
            - email:
                is: user@example.com
            - groups:
                has: "admins"
            - groups:
                has: "developers"
  - from: https://hello.localhost.pomerium.io
    to: http://localhost:8080
    policy:
      - allow:
          or:
            - groups:
                has: "admins@pomerium.io"

# CORS Preflight

  • yaml/json setting: cors_allow_preflight
  • Type: bool
  • Optional
  • Default: false

Allow unauthenticated HTTP OPTIONS requests as per the CORS spec (opens new window).

# Enable Google Cloud Serverless Authentication

  • Environmental Variable: ENABLE_GOOGLE_CLOUD_SERVERLESS_AUTHENTICATION
  • Config File Key: enable_google_cloud_serverless_authentication
  • Type: bool
  • Default: false

Enable sending a signed Authorization Header (opens new window) to upstream GCP services.

Requires setting Google Cloud Serverless Authentication Service Account or running Pomerium in an environment with a GCP service account present in default locations.

# From

  • yaml/json setting: from
  • Type: URL (must contain a scheme and hostname, must not contain a path)
  • Schemes: https, tcp+https
  • Required
  • Example: https://verify.corp.example.com, tcp+https://ssh.corp.example.com:22

From is the externally accessible URL for the proxied request.

Specifying tcp+https for the scheme enables TCP proxying support for the route. You may map more than one port through the same hostname by specifying a different :port in the URL.

WARNING

Only secure schemes (https and tcp+https) are supported.

# Kubernetes Service Account Token

  • yaml/json setting: kubernetes_service_account_token / kubernetes_service_account_token_file
  • Type: string or relative file location containing a Kubernetes bearer token
  • Optional
  • Example: eyJ0eXAiOiJKV1QiLCJhbGciOiJ... or /var/run/secrets/kubernetes.io/serviceaccount/token

Use this token to authenticate requests to a Kubernetes API server.

Pomerium will impersonate (opens new window) the Pomerium user's identity, and Kubernetes RBAC can be applied to IdP user and groups.

# Signout Redirect URL

  • Environmental Variable: SIGNOUT_REDIRECT_URL
  • Config File Key: signout_redirect_url
  • Type: URL
  • Required
  • Example: https://signout-redirect-url.corp.example.com

Signout redirect url is the url user will be redirected to after signing out.

You can overwrite this behavior by passing the query param pomerium_redirect_uri or post value pomerium_redirect_uri to the /.pomerium/signout/ endpoint.

# Path

  • yaml/json setting: path
  • Type: string
  • Optional
  • Example: /admin/some/exact/path

If set, the route will only match incoming requests with a path that is an exact match for the specified path.

# Prefix

  • yaml/json setting: prefix
  • Type: string
  • Optional
  • Example: /admin

If set, the route will only match incoming requests with a path that begins with the specified prefix.

# Prefix Rewrite

  • yaml/json setting: prefix_rewrite
  • Type: string
  • Optional
  • Example: /subpath

If set, indicates that during forwarding, the matched prefix (or path) should be swapped with this value. For example, given this policy:

from: https://from.example.com
to: https://to.example.com
prefix: /admin
prefix_rewrite: /

A request to https://from.example.com/admin would be forwarded to https://to.example.com/.

# Host Rewrite

  • yaml/json settings: host_rewrite, host_rewrite_header, host_path_regex_rewrite_pattern, host_path_regex_rewrite_substitution
  • Type: string
  • Optional
  • Example: host_rewrite: "example.com"

The host header can be preserved via the preserve_host_header setting or customized via 3 mutually exclusive options:

  1. preserve_host_header when enabled, this option will pass the host header from the incoming request to the proxied host, instead of the destination hostname. It's an optional parameter of type bool that defaults to false. See ProxyPreserveHost (opens new window).

  2. host_rewrite which will rewrite the host to a new literal value.

  3. host_rewrite_header which will rewrite the host to match an incoming header value.

  4. host_path_regex_rewrite_pattern, host_path_regex_rewrite_substitution which will rewrite the host according to a regex matching the path. For example with the following config:

    host_path_regex_rewrite_pattern: "^/(.+)/.+$"
host_path_regex_rewrite_substitution: \1

    Would rewrite the host header to example.com given the path /example.com/some/path.

The 2nd, 3rd and 4th options correspond to the envoy route action host related options, which can be found here (opens new window).

# Public Access

  • yaml/json setting: allow_public_unauthenticated_access
  • Type: bool
  • Optional
  • Default: false

Use with caution: Allow all requests for a given route, bypassing authentication and authorization. Suitable for publicly exposed web services.

If this setting is enabled, no whitelists (e.g. Allowed Users) should be provided in this route.

# Allow Any Authenticated User

  • yaml/json setting: allow_any_authenticated_user
  • Type: bool
  • Optional
  • Default: false

Use with caution: This setting will allow all requests for any user which is able to authenticate with our given identity provider. For instance, if you are using a corporate GSuite account, an unrelated gmail user will be able to access the underlying upstream.

Use of this setting means Pomerium will not enforce centralized authorization policy for this route. The upstream is responsible for handling any authorization.

# Regex

  • yaml/json setting: regex
  • Type: string (containing a regular expression)
  • Optional
  • Example: ^/(admin|superuser)/.*$

If set, the route will only match incoming requests with a path that matches the specified regular expression. The supported syntax is the same as the Go regexp package (opens new window) which is based on re2 (opens new window).

# Regex Rewrite

  • yaml/json setting: regex_rewrite_pattern, regex_rewrite_substitution
  • Type: string
  • Optional
  • Example: { "regex_rewrite_pattern":"^/service/([^/]+)(/.*)$", "regex_rewrite_substitution": "\\2/instance/\\1" }

If set, the URL path will be rewritten according to the pattern and substitution, similar to prefix_rewrite.

# Outlier Detection

  • yaml/json setting: outlier_detection
  • Type: object
  • Optional
  • Example: { "consecutive_5xx": 12 }

Outlier detection and ejection is the process of dynamically determining whether some number of hosts in an upstream cluster are performing unlike the others and removing them from the healthy load balancing set.

See Envoy documentation (opens new window) and API (opens new window) for more details.

# Route Timeout

Policy timeout establishes the per-route timeout value. Cannot exceed global timeout values.

# Idle Timeout

If you are proxying long-lived requests that employ streaming calls such as websockets or gRPC, set this to either a maximum value there may be no data exchange over a connection (recommended), or set it to unlimited (0s). If idle_timeout is specified, and timeout is not explicitly set, then timeout would be unlimited (0s). You still may specify maximum lifetime of the connection using timeout value (i.e. to 1 day).

# Set Request Headers

  • Config File Key: set_request_headers
  • Type: map of strings key value pairs
  • Optional

Set Request Headers allows you to set static values for given request headers. This can be useful if you want to pass along additional information to downstream applications as headers, or set authentication header to the request. For example:

- from: https://verify.corp.example.com
  to: https://verify.pomerium.com
  policy:
    - allow:
        or:
          - email:
              is: user@example.com
  set_request_headers:
    # works auto-magically!
    # https://verify.corp.example.com/basic-auth/root/hunter42
    Authorization: Basic cm9vdDpodW50ZXI0Mg==
    X-Your-favorite-authenticating-Proxy: "Pomerium"

WARNING

Neither :-prefixed pseudo-headers nor the Host: header may be modified via this mechanism. Those headers may instead be modified via mechanisms such as prefix_rewrite, regex_rewrite, and host_rewrite.

# Remove Request Headers

  • Config File Key: remove_request_headers
  • Type: array of strings
  • Optional

Remove Request Headers allows you to remove given request headers. This can be useful if you want to prevent privacy information from being passed to downstream applications. For example:

- from: https://verify.corp.example.com
  to: https://verify.pomerium.com
  policy:
    - allow:
        or:
          - email:
              is: user@example.com
  remove_request_headers:
    - X-Email
    - X-Username

# Set Response Headers

  • Config File Key: set_response_headers
  • Type: map of strings key value pairs
  • Optional

Set Response Headers allows you to set static values for the given response headers. These headers will take precedence over the global set_response_headers.

# Rewrite Response Headers

  • Config File Key: rewrite_response_headers
  • Type: object
  • Optional
  • Example: [{ "header": "Location", "prefix": "http://localhost:8000/two/", "value": "http://frontend/one/" }]

Rewrite Response Headers allows you to modify response headers before they are returned to the client. The header field will match the HTTP header name, and prefix will be replaced with value. For example, if the downstream server returns a header:

Location: http://localhost:8000/two/some/path/

And the policy has this config:

rewrite_response_headers:
  - header: Location
    prefix: http://localhost:8000/two/
    value: http://frontend/one/

The browser would be redirected to: http://frontend/one/some/path/. This is similar to nginx's proxy_redirect option (opens new window), but can be used for any header.

# Redirect

  • yaml/json setting: 'redirect'
  • Type: object
  • Optional
  • Example: { "host_redirect": "example.com" }

Redirect is used to redirect incoming requests to a new URL. The redirect field is an object with several possible options:

  • https_redirect (boolean): the incoming scheme will be swapped with "https".
  • scheme_redirect (string): the incoming scheme will be swapped with the given value.
  • host_redirect (string): the incoming host will be swapped with the given value.
  • port_redirect (integer): the incoming port will be swapped with the given value.
  • path_redirect (string): the incoming path portion of the URL will be swapped with the given value.
  • prefix_rewrite (string): the incoming matched prefix will be swapped with the given value.
  • response_code (integer): the response code to use for the redirect. Defaults to 301.
  • strip_query (boolean): indicates that during redirection, the query portion of the URL will be removed. Defaults to false.

Either redirect or to must be set.

# To

  • yaml/json setting: to
  • Type: URL or list of URLs (must contain a scheme and hostname) with an optional weight
  • Schemes: http, https, tcp
  • Optional
  • Example: http://verify , https://192.1.20.12:8080, http://neverssl.com, https://verify.pomerium.com/anything/, ["http://a", "http://b"], ["http://a,10", "http://b,20"]

To is the destination(s) of a proxied request. It can be an internal resource, or an external resource. Multiple upstream resources can be targeted by using a list instead of a single URL:

- from: https://example.com
  to:
  - https://a.example.com
  - https://b.example.com

A load balancing weight may be associated with a particular upstream by appending ,[weight] to the URL. The exact behavior depends on your lb_policy setting. See Load Balancing for example configurations.

Must be tcp if from is tcp+https.

WARNING

Be careful with trailing slash.

With rule:

- from: https://verify.corp.example.com
  to: https://verify.pomerium.com/anything

Requests to https://verify.corp.example.com will be forwarded to https://verify.pomerium.com/anything, while requests to https://verify.corp.example.com/foo will be forwarded to https://verify.pomerium.com/anythingfoo.To make the request forwarded to https://httbin.org/anything/foo, you can use double slashes in your request https://httbin.corp.example.com//foo.

While the rule:

- from: https://verify.corp.example.com
  to: https://verify.pomerium.com/anything/

All requests to https://verify.corp.example.com/* will be forwarded to https://verify.pomerium.com/anything/*. That means accessing to https://verify.corp.example.com will be forwarded to https://verify.pomerium.com/anything/. That said, if your application does not handle trailing slash, the request will end up with 404 not found.

Either redirect or to must be set.

# TLS Skip Verification

  • Config File Key: tls_skip_verify
  • Type: bool
  • Default: false

TLS Skip Verification controls whether a client verifies the server's certificate chain and host name. If enabled, TLS accepts any certificate presented by the server and any host name in that certificate. In this mode, TLS is susceptible to man-in-the-middle attacks. This should be used only for testing.

# TLS Server Name

  • Config File Key: tls_server_name
  • Type: string
  • Optional

TLS Server Name overrides the hostname specified in the to field. If set, this server name will be used to verify the certificate name. This is useful when the backend of your service is an TLS server with a valid certificate, but mismatched name.

# TLS Custom Certificate Authority

TLS Custom Certificate Authority defines a set of root certificate authorities that clients use when verifying server certificates.

Note: This setting will replace (not append) the system's trust store for a given route.

# TLS Downstream Client Certificate Authority

If specified downstream clients (eg a user's browser) will be required to provide a valid client TLS certificate. This overrides the global client_ca option for this route.

# TLS Client Certificate

  • Config File Key: tls_client_cert and tls_client_key or tls_client_cert_file and tls_client_key_file
  • Type: base64 encoded (opens new window) string or relative file location
  • Optional

Pomerium supports client certificates which can be used to enforce mutually authenticated and encrypted TLS connections (opens new window) (mTLS). For more details, see our mTLS example repository (opens new window) and the certificate docs.

# Pass Identity Headers

  • yaml/json setting: pass_identity_headers
  • Type: bool
  • Optional
  • Default: false

When enabled, this option will pass identity headers to upstream applications. These headers include:

  • X-Pomerium-Jwt-Assertion
  • X-Pomerium-Claim-*

# SPDY

  • Config File Key: allow_spdy
  • Type: bool
  • Default: false

If set, enables proxying of SPDY protocol upgrades.

# Cluster Name

  • Config File Key: name
  • Type: string
  • Optional

Runtime metrics for this policy would be available under envoy_cluster_name prefix.

# Load Balancing Policy

  • Config File Key: lb_policy
  • Type: enum
  • Optional

In presence of multiple upstreams, defines load balancing strategy between them.

See Envoy documentation (opens new window) for more details.

Some policy types support additional configuration.

# Load Balancing Policy Config

  • Config File Key: least_request_lb_config, ring_hash_lb_config, maglev_lb_config
  • Type: object
  • Optional

When lb_policy is configured, you may further customize policy settings for LEAST_REQUEST, RING_HASH, AND MAGLEV using one of the following options.

See Load Balancing for example configurations

# Health Checks

  • Config File Key: health_checks
  • Type: array of objects
  • Optional

When defined, will issue periodic health check requests to upstream servers. When health checks are defined, unhealthy upstream servers would not serve traffic. See also outlier_detection for automatic upstream server health detection. In presence of multiple upstream servers, it is recommended to set up either health_checks or outlier_detection or both.

See Envoy documentation (opens new window) for a list of supported parameters (opens new window).

Only one of http_health_check, tcp_health_check, or grpc_health_check may be configured per health_check object definition.

See Load Balancing for example configurations.

# Websocket Connections

  • Config File Key: allow_websockets
  • Type: bool
  • Default: false

If set, enables proxying of websocket connections.

WARNING

Use with caution: websockets are long-lived connections, so global timeouts are not enforced (though the policy-specific timeout is enforced). Allowing websocket connections to the proxy could result in abuse via DOS attacks (opens new window).

# Authorize Service

# Authorize Service URL

  • Environmental Variable: AUTHORIZE_SERVICE_URL or AUTHORIZE_SERVICE_URLS
  • Config File Key: authorize_service_url or authorize_service_urls
  • Type: URL
  • Required
  • Example: https://authorize.corp.example.com

Authorize Service URL is the location of the internally accessible Authorize service. Multiple URLs can be specified with authorize_service_urls.

# Google Cloud Serverless Authentication Service Account

  • Environmental Variable: GOOGLE_CLOUD_SERVERLESS_AUTHENTICATION_SERVICE_ACCOUNT
  • Config File Key: google_cloud_serverless_authentication_service_account
  • Type: base64 encoded (opens new window) string
  • Optional

Manually specify the service account credentials to support GCP's Authorization Header (opens new window) format.

If unspecified:

# Signing Key

Signing Key is the private key used to sign a user's attestation JWT which can be consumed by upstream applications to pass along identifying user information like username, id, and groups.

If set, the signing key's public key will can retrieved by hitting Pomerium's /.well-known/pomerium/jwks.json endpoint which lives on the authenticate service. Otherwise, the endpoint will return an empty keyset.

For example, assuming you have generated an ES256 key (opens new window) as follows.

# Generates an P-256 (ES256) signing key
openssl ecparam  -genkey  -name prime256v1  -noout  -out ec_private.pem
# careful! this will output your private key in terminal
cat ec_private.pem | base64

That signing key can be accessed via the well-known jwks endpoint.

$ curl https://authenticate.int.example.com/.well-known/pomerium/jwks.json | jq

{
  "keys": [
    {
      "use": "sig",
      "kty": "EC",
      "kid": "ccc5bc9d835ff3c8f7075ed4a7510159cf440fd7bf7b517b5caeb1fa419ee6a1",
      "crv": "P-256",
      "alg": "ES256",
      "x": "QCN7adG2AmIK3UdHJvVJkldsUc6XeBRz83Z4rXX8Va4",
      "y": "PI95b-ary66nrvA55TpaiWADq8b3O1CYIbvjqIHpXCY"
    }
  ]
}

If no certificate is specified, one will be generated and the base64'd public key will be added to the logs. Note, however, that this key be unique to each service, ephemeral, and will not be accessible via the authenticate service's jwks_uri endpoint.