OpenTelemetry web server metrics reference

Platform 6.6.0+

Itential Platform provides the GET /metrics/webserver endpoint, which returns metrics about its web server in Prometheus text format. The endpoint requires an authenticated session. Monitoring tools such as Prometheus, Grafana, and Datadog can use the endpoint to track web server health.

This endpoint is separate from the /prometheus_metrics route, which reports a different set of Platform metrics. You can use both.

Metrics

The endpoint reports only on web server health.

Metric group	Description	Broken out by
Request totals	The number of API requests the web server handled.	HTTP method, status code class (2xx, 3xx, 4xx, or 5xx), and caller type
Latency	Request and response duration, reported as a histogram.	Not applicable
Errors	The number of requests that returned an error.	Route and status code, such as 401, 403, 404, or 500
Payload size	The size of request and response payloads.	Not applicable

Metric behavior

Values are aggregate totals since the process last started. Totals reset to zero when Platform restarts.
The endpoint reports current values only. It doesn’t store history. To view trends over time, use a monitoring tool that records the values at regular intervals.

Configuration

Web server metrics collection is enabled by default. The following Platform property and environment variable turn it on or off. For all Platform properties, see the properties and environment variables reference.

Property	Environment variable	Type	Default
`otel_webserver_metrics_enabled`	`ITENTIAL_OTEL_WEBSERVER_METRICS_ENABLED`	Boolean	`true`

Limitations

Itential doesn’t report the following metrics:

Node.js and V8 runtime internals, such as event loop lag and garbage collection.
Per-process CPU and memory breakdowns.
Time in queue.
Workflow engine and general server CPU and memory metrics. To monitor them, use exporters.