finds.dev← search

// the find

sibson/redbeat

★ 1,045 · Python · Apache-2.0 · updated Mar 2026

RedBeat is a Celery Beat Scheduler that stores the scheduled tasks and runtime metadata in Redis.

RedBeat replaces Celery Beat's default file-based scheduler with Redis, meaning your periodic task definitions survive restarts, can be modified at runtime without redeploying, and don't break when you run multiple instances. It's for anyone running Celery at more than one-box scale who has hit the pain of Beat being a single-file, single-machine bottleneck.

The distributed lock via Redis prevents the classic 'accidentally ran two Beat processes and now everything fired twice' failure mode. Tasks can be added, modified, or deleted at runtime through any Redis client — no redeploy, no restart. Startup with large task counts is fast because it only loads the next-due entries rather than the full schedule. The design doc in `docs/design.rst` actually explains the data model, which is rare and useful when you need to debug what's in Redis.

Redis becomes a hard dependency for your scheduler — if Redis goes down, Beat stops entirely. With the default file-based scheduler you at least have a degraded but running system. The codebase is compact (four source files), which is fine, but it means there's essentially no abstraction layer if you need to swap Redis for something else later. Documentation covers config but doesn't explain what happens during Redis failover or whether Sentinel/Cluster setups are fully supported — you'll find out at 2am. Last substantive activity looks sparse; 1k stars over several years suggests it works but isn't being actively developed.

View on GitHub →

// want more like this?

We dig through GitHub every week and send a few repos picked for what you actually care about — each with an honest take like this one.

Get finds in your inbox → Search again →