Boundary metrics

Boundary provides metrics through the /metrics path using a listener with the "ops" purpose. By default, a listener with that purpose runs on port 9203. See the example configuration section for an example listener stanza in a config.hcl file.

Metrics are important for understanding what is happening in a Boundary deployment, enabling diagnoses of deployed controllers and workers. The exported metrics grant greater visibility into the various components of a running Boundary deployment through Prometheus.

The metrics uses the OpenMetric exposition format which can be read by Prometheus.

Controller

The following table lists the current names and descriptions of metrics provided by controllers.

Name	Description
`boundary_controller_api_http_request_duration_seconds`	Histogram of latencies for HTTP requests.
`boundary_controller_api_http_request_size_bytes`	Histogram of request sizes for HTTP requests.
`boundary_controller_api_http_response_size_bytes`	Histogram of response sizes for HTTP requests.
`boundary_controller_cluster_grpc_request_duration_seconds`	Histogram of latencies for requests made to the gRPC service running on the cluster listener.

Worker

The following table lists the current names and descriptions of metrics provided by workers.

Name	Description
`boundary_worker_proxy_http_write_header_duration_seconds`	Histogram of time elapsed after the TLS connection is established to when the first http header is written back from the server.
`boundary_worker_proxy_websocket_active_connections`	A gauge of the current count of open proxy connections on the worker.
`boundary_worker_proxy_websocket_received_bytes_total`	Count of received bytes sent over all proxy connections handled by the worker.
`boundary_worker_proxy_websocket_sent_bytes_total`	Count of sent bytes sent over all proxy connections handled by the worker.

Other

The following table lists the current miscellaneous metrics provided by Boundary.

Name	Description
`boundary_build_info`	A gauge set to 1 with label values detailing the current Go version, git revision, and Boundary version.

Metric labels

Each metric uses labels to help organize and narrow its scope so that users can filter for different characteristics, allowing more flexible analysis of the measurements collected.

Metrics for HTTP requests include the following labels:

Label	Description
`code`	The HTTP status code of the request (e.g., “200”, “400”, etc.).
`path`	The path, excluding query parameters, associated with the request. The value for this label replaces resource IDs with "{id}" (for example, `/v1/groups/{id}` is a possible value for this label). Any unexpected paths receive the value "unknown."
`method`	The name of the method on the HTTP request. For example "GET", “POST”, “DELETE”.

Metrics for gRPC requests include the following labels:

Label	Description
`grpc_method`	The name of the method on the grpc service (e.g., `GetGroup`).
`grpc_service`	The proto service name including the package (e.g., `controller.api.services.v1.GroupService`).
`grpc_code`	The grpc status code in human-readable format. For example, `OK`, `IllegalArgument`, `Unknown`.

Example configuration

Defining a listener stanza in the config file is sufficient for enabling metrics reporting in Boundary. For details on what fields are allowed in this stanza, refer to the documentation about TCP Listener.

An example listener stanza:

listener "tcp" {
    purpose = "ops"
    tls_disable = true
}

Tutorial

Refer to the Prometheus Metrics tutorial for a step-by-step introduction to metrics.