Celery

This is the standard for async task queues in Python, and for good reason. It handles the whole distributed job workflow: broker setup with Redis or RabbitMQ, worker management, retries with exponential backoff, periodic scheduling, and task chaining. The progressive disclosure here walks through real broker configs, task routing patterns, and production concerns like monitoring and rate limiting. Use it when you need to offload email sending, report generation, or any long-running work that would block a web request. The overhead is real though. If you just need simple background jobs and don't care about distribution, RQ or even basic asyncio might be easier to reason about.

Install to Claude Code

npx -y skills add bobmatnyc/claude-mpm-skills --skill celery --agent claude-code