Metrics | KEDA

Metrics

Reference for all Prometheus metrics exposed by the HTTP Add-on

Warning

You are currently viewing v0.16 of the documentation and it is not the latest. For the most recent documentation, kindly click here.

The interceptor and the external scaler expose metrics via OpenTelemetry, with a Prometheus-compatible endpoint enabled by default on each component.

Endpoint configuration

Setting	Default	Description
Port	`2223`	Configurable via `OTEL_PROM_EXPORTER_PORT`.
Path	`/metrics`	Standard Prometheus scrape path.
Enabled	`true`	Configurable via `OTEL_PROM_EXPORTER_ENABLED`.

Both components also support OTLP metric export. See Environment Variables for configuration.

Interceptor metrics

Request count


Prometheus name	`interceptor_request_count_total`
OTel instrument name	`interceptor.request.count`
Type	Counter
Description	Total requests processed by the interceptor proxy.

Labels:

Label	Description
`code`	HTTP response status code (integer).
`cold_start`	`true` when the request waited for backend readiness, including waits that ended before the backend became ready; otherwise `false`.
`method`	HTTP request method. Non-standard methods are normalized to `_OTHER` (see Method normalization).
`route_name`	Name of the matched InterceptorRoute or HTTPScaledObject.
`route_namespace`	Namespace of the matched route resource.

Request concurrency


Prometheus name	`interceptor_request_concurrency`
OTel instrument name	`interceptor.request.concurrency`
Type	UpDownCounter (gauge)
Description	Requests currently in-flight at the interceptor proxy.

Labels:

Label	Description
`route_name`	Name of the matched InterceptorRoute or HTTPScaledObject.
`route_namespace`	Namespace of the matched route resource.

Request duration


Prometheus name	`interceptor_request_duration_seconds`
OTel instrument name	`interceptor.request.duration`
Type	Histogram
Unit	Seconds
Description	Time from request received to response written.

Bucket boundaries: 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5, 7.5, 10, 15, 30, 60, 120, 300. The first boundaries follow the OTel HTTP semantic conventions, and the additional boundaries expose longer cold-start request durations.

Labels:

Label	Description
`code`	HTTP response status code (integer).
`cold_start`	`true` when the request waited for backend readiness, including waits that ended before the backend became ready; otherwise `false`.
`method`	HTTP request method. Non-standard methods are normalized to `_OTHER` (see Method normalization).
`route_name`	Name of the matched InterceptorRoute or HTTPScaledObject.
`route_namespace`	Namespace of the matched route resource.

Cold-start duration


Prometheus name	`interceptor_cold_start_duration_seconds`
OTel instrument name	`interceptor.cold_start.duration`
Type	Histogram
Unit	Seconds
Description	Time spent waiting for a cold-start backend to become ready.

The metric is recorded only when a request waits for backend readiness.

Bucket boundaries: 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5, 7.5, 10, 15, 30, 60, 120, 300.

Labels:

Label	Description
`outcome`	Readiness wait result: `ready` when backend became ready, `timeout` when wait ended with another error, or `cancelled` when request context was cancelled.
`route_name`	Name of the matched InterceptorRoute or HTTPScaledObject.
`route_namespace`	Namespace of the matched route resource.

Cold-start rejections


Prometheus name	`interceptor_cold_start_rejections_total`
OTel instrument name	`interceptor.cold_start.rejections`
Type	Counter
Description	Requests rejected because the cold-start pending request limit was reached.

Labels:

Label	Description
`route_name`	Name of the matched InterceptorRoute or HTTPScaledObject.
`route_namespace`	Namespace of the matched route resource.

Method normalization

The method label accepts the following standard HTTP methods without modification:

CONNECT, DELETE, GET, HEAD, OPTIONS, PATCH, POST, PUT, TRACE

All other method values are replaced with _OTHER to prevent unbounded label cardinality.

Scaler metrics

Pinger fetch duration


Prometheus name	`scaler_pinger_fetch_duration_seconds`
OTel instrument name	`scaler.pinger.fetch.duration`
Type	Histogram
Unit	Seconds
Description	Duration of a queue pinger fetch cycle across all interceptor pods.

Bucket boundaries: 0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1, 2.5, 5

Pinger fetch errors


Prometheus name	`scaler_pinger_fetch_errors_total`
OTel instrument name	`scaler.pinger.fetch.errors`
Type	Counter
Description	Total failed queue pinger fetch cycles.

Pinger endpoints


Prometheus name	`scaler_pinger_endpoints`
OTel instrument name	`scaler.pinger.endpoints`
Type	Gauge
Description	Number of interceptor endpoints the scaler is polling.