Demo stack¶

Screenshots from manage.py o11y stack start running against the test project, a minimal Django app that fires one Celery task every 5 seconds via a /trigger/ endpoint. Traffic is low so most panels are sparse, but everything is wired up.

Run it yourself with:

python manage.py o11y stack start

Grafana starts at http://localhost:3000 with all dashboards pre-loaded and no login required.

Metrics¶

Django request rates, latency percentiles, database query counts, cache hit rates, and migration status from django-prometheus. Dashboards imported from django-mixin.

Metrics dashboard

Traces¶

Distributed traces from OpenTelemetry via Tempo. Each HTTP request and Celery task gets a span, with database queries and outbound HTTP calls as child spans.

Traces dashboard

Logs¶

Structured JSON logs from structlog ingested via Alloy into Loki. Every log line carries trace_id and span_id, so you can jump from a log entry directly to its trace.

Logs dashboard

Profiles¶

Continuous CPU profiling from Pyroscope. Flame graphs show where your app spends time. Profiles carry the same service.name and deployment.environment tags as your traces.

Profiles dashboard

Imported dashboards¶

These dashboards are pulled from Grafana.com when you run o11y stack start. They come from django-mixin and celery-mixin. If a panel is wrong or a dashboard is missing something, open an issue in the relevant mixin repo rather than here.