// the find
xuxueli/xxl-job
A distributed task scheduling framework.（分布式任务调度平台XXL-JOB）
XXL-JOB is a distributed task scheduling platform for Java backends — think Quartz with a proper admin UI, HA clustering, and executor auto-registration baked in. It's been running in production at hundreds of Chinese companies since 2015, which is about as real-world-tested as it gets. The target is teams that need cron/event/interval scheduling across a cluster without wiring it all together themselves.
The scheduling center uses DB-level locking to guarantee exactly-once triggering across a cluster, which is the hard part that most homegrown Quartz setups get wrong. The executor routing options are genuinely useful — consistent hash, LRU, LFU, failover, and busy-transfer all solve real problems depending on your job shape. GLUE mode (live code editing via web IDE with 30-version history) is a practical escape hatch for ops teams who need to hotfix job logic without a deployment. The thread pool isolation between normal and slow jobs prevents a handful of long-running tasks from starving the scheduler.
The license is GPL v3, which is a hard stop for any proprietary product that wants to embed or modify it — you'd need to negotiate a commercial license or fork it and stay upstream. Documentation outside of Chinese is thin; the English docs exist but lag behind the Chinese version and miss details you'll need. The admin UI is built on AdminLTE/Bootstrap 3, which is dated and will feel clunky if your team is used to modern tooling. Security posture is basic — two hardcoded user roles with executor-level ACL, no RBAC, no audit log beyond the recent addition, and the access token model for executor-to-scheduler auth is a single shared secret per executor group.