Prometheus metrics

When metrics_listen is set, Qrux serves Prometheus text format at /metrics on that address.

Metric names use the qrux_ prefix.

Example:

http://127.0.0.1:9090/metrics

Typical metrics

Request counts (labels: method, status, upstream):

qrux_http_requests_total{method="GET",status="200",upstream="example.com:80"} 42

Latency (histogram, labels: method, upstream):

qrux_http_request_duration_seconds_bucket{method="GET",upstream="example.com:80",le="0.1"} 40

Active QUIC connections:

qrux_active_connections 5

Pooled upstream TCP connections (label: upstream):

qrux_upstream_pool_connections{upstream="example.com:80"} 3

Upstream request timeouts (whole request exceeded upstream_request_timeout_secs):

qrux_upstream_timeouts_total 2

Scrape this endpoint with Prometheus or inspect it with curl while debugging.