Spans for everything, with no code to write.

Turn on the telemetry feature and every request, query, cache call, job, mail and event opens a span, exported over OTLP to your collector.

Read the Observability docs

telemetry example output: spans for the HTTP request, cache, query, outbound HTTP call and queue, then queue metrics.

What you get

The standard OTEL_* variables, so an existing deployment’s settings work.
Prometheus metrics by route pattern, at a path you choose.
Health checks: Health::new().register("database", …).
A debug error page only when APP_DEBUG=true — and it refuses to render in production.
The same spans feed Lookout, in-process.

In the docs

Observability
Lookout

.env.production

# Spans for every request, query, cache call, job, mail and event —
# exported over OTLP. No application code opens one.
TELEMETRY_ENABLED=true
OTEL_EXPORTER_OTLP_ENDPOINT=http://otel-collector:4317

# The same spans, kept in-process and shown at /_lookout.
LOOKOUT_ENABLED=true
LOOKOUT_STORE=cache

Up close

Select an image to see it full size.

The debug error page for a 500: the message, frame filters and the stack frames. — With APP_DEBUG=true and APP_ENV=local, a 500 shows the stack, filterable to app, framework or vendor frames; in production it refuses to render.

telemetry example output: an outbound GET carrying a traceparent header with the request trace id. — The outbound HTTP client propagates the trace, so the service on the other end joins the same trace.

Prometheus-style queue metrics: job duration, wait time, jobs in flight and jobs total. — The worker records job duration, wait time and throughput per job and queue, with no application code.

curl output of /health and /health/version JSON. — Health and build-version endpoints for probes and deploy checks.

All features