Observability

Use these pages to monitor Smart Router in production and find out what it is doing. They cover four topics:

Topic	What's in it
Metrics	Every Prometheus metric exposed, with types and labels, plus PromQL recipes and suggested alerts.
Traces	OpenTelemetry spans over the full relay lifecycle, configured with standard OTel env vars.
Profiling	Continuous CPU / memory / mutex / block profiles to Pyroscope (or one-off pprof).
Error codes	The internal error taxonomy seen in logs and the smartrouter_errors_total metric, with retryability.

The prebuilt dashboard overlay graphs the metric series with no extra setup. Most signals are off until you connect the router to a collector, Pyroscope, or Prometheus. Each of these pages explains how to turn its signal on.