// the take
Overlord is Bilibili's internal cache proxy and cluster management platform for Redis and Memcache, written in Go. It sits in the same space as Twemproxy but adds redis-cluster support and a web dashboard for cluster lifecycle management (create, scale, delete nodes). Built for a large production environment and used at Bilibili's scale.
The proxy's ability to present itself as a redis-cluster node to clients is genuinely useful — it lets you swap backends without changing client config. The fuzz test corpus for both the memcache and Redis parsers shows they took protocol correctness seriously, which matters for a proxy that touches every cache request. The slot migration tooling in enri is practical: slot rebalancing during scale-out is usually the painful manual step in Redis cluster operations. The architecture is cleanly split — proxy, platform orchestrator, sync tool, and cluster manager are separate binaries rather than one monolith.
The whole platform layer is built on Apache Mesos, which is effectively dead infrastructure at this point — if you're not already running Mesos, this part of the stack is unusable without a significant rewrite. Last commit was July 2023 and there's been no meaningful activity since; Redis 7.x and recent protocol changes are almost certainly not covered. All documentation is in Chinese, which is fine for Bilibili's team but a real barrier for anyone outside that context. The web GUI is a Vue 2 app with its own separate dependency tree and no clear path to integrate or deploy outside the full platform stack.
We read the README, the directory tree and a key source file in July 2026. Nobody pays to be listed. How we judge →