Observability

Logs, metrics and traces of a running service: each comes from a layer on the app and a feature that exports it.

RustStream reports what it does through tracing events, and the logging feature prints them on the console. The per-delivery signals come from layers that wrap every handler, so a handler body holds only its own logic. A layer added to the app wraps the handlers of its routers too. Prometheus metrics need the metrics feature; trace propagation and OpenTelemetry export need the otel feature.