Monitoring
Talos exposes Prometheus metrics, OpenTelemetry tracing, and health endpoints.
- Prometheus metrics — request counts, latencies, and pool sizes
- OpenTelemetry tracing — distributed request traces
- Health checks — liveness and readiness probes
Edition differences
Both editions record business-logic metrics in-process: counters for keys created, verified, revoked, and rotated, plus cache hit and miss counters. The editions differ in what they expose and which signals they collect:
| Feature | OSS | Commercial |
|---|---|---|
| In-process metric collection | Yes | Yes |
/metrics scrape endpoint (port 4422) | No | Yes |
/health/alive and /health/ready (main port 4420) | Yes | Yes |
/health/alive and /health/ready (metrics port 4422) | Yes | Yes |
OpenTelemetry trace export (otlp exporter) | No | Yes |
The OSS edition doesn't register a /metrics route or export traces. To scrape metrics or ship traces to a collector, use the
commercial edition. Health endpoints work in both editions on both ports.
