Metrics & Monitoring

The API exposes HTTP request metrics for Prometheus when enabled. Full config, runbook, and PromQL: Observability overview — Metrics.

Config

Variable	Default	Description
METRICS_ENABLED	false	Expose GET /metrics
METRICS_ENDPOINT	/metrics	Path for the endpoint
METRICS_COLLECT_DEFAULTS	true	CPU, memory, event loop

Defined in apps/api/src/config/env.config.ts; implementation in apps/api/src/lib/metrics/.

Running the stack

See observability/README.md in the repo:

1. Start the API with METRICS_ENABLED=true (and APP_PORT=4000 or match observability/prometheus.yml).
2. Run docker compose up -d prometheus grafana.
3. In Grafana (e.g. http://localhost:3001), add a Prometheus data source with URL http://prometheus:9090.

Example PromQL

Request rate:

rate(http_requests_total[5m])

95th percentile duration by route:

histogram_quantile(0.95, sum(rate(http_request_duration_seconds_bucket[5m])) by (le, route))

Full metrics section: Observability overview — Metrics.

---

Related:

• Observability overview — Logging, metrics, telemetry
• Configuration — Environment variable reference