Why it exists
Any project running Celery eventually needs a way to see what's queued, running, and failing without grepping worker logs. This packages Flower as a standalone image that points at an existing broker, so it drops into a docker-compose stack alongside the workers it's watching.
What I learned
Observability tooling is worth setting up before the queue backs up,
not after. Having Flower one docker run away made a few otherwise
confusing "why isn't this task running" moments trivial to diagnose.
Status
Active, used on projects that run a Celery/RabbitMQ queue.