// the find
mher/flower
Real-time monitor and web admin for Celery distributed task queue
Flower is a web dashboard and REST API for monitoring Celery task queues in real time. It shows worker status, task history, queue depths, and lets you issue remote control commands against running workers. If you run Celery in production and have ever needed to know why a task is stuck, this is the tool.
Real-time event-driven updates via Celery's own event bus — no polling hack, no separate agent to deploy. Remote control is genuinely useful: you can resize worker pools, revoke tasks, and restart workers without SSH. Prometheus integration means you can bolt this onto an existing Grafana stack and get queue-depth alerts without building anything. Auth options cover Basic, OAuth (Google/GitHub/GitLab/Okta), so you're not stuck bolting on a reverse proxy just to lock it down.
Task history is in-memory by default and lost on restart — there's a persistence option but it requires extra config and it's not obvious from the docs. The UI is jQuery + DataTables from 2023, which isn't a dealbreaker but shows this isn't getting active frontend investment. No multi-tenant support: one Flower instance sees one broker, so multi-environment setups mean running multiple Flower processes. Issue tracker has a long tail of open bugs that don't get closed; the maintainer is responsive in bursts but this is essentially a one-person project, which is a real adoption risk for anything you'd call production-critical infrastructure.