// the find
lyft/clutch
Extensible platform for infrastructure management
Clutch is Lyft's internal developer portal for infrastructure management — a Go+React platform where engineering teams build workflows for things like resizing EC2 instances, managing Kubernetes deployments, or running chaos experiments, all behind SSO and RBAC. It's for platform/infra teams who want a standardized UI for ops tasks instead of giving everyone raw AWS console access. The bad news is in the README: Lyft has announced it's being archived.
The Resolver pattern is genuinely clever — server-generated forms from Protobuf definitions mean your backend and frontend stay in sync without maintaining two schema definitions. The single-binary deployment (frontend assets embedded in the Go binary) removes a whole class of deployment coordination problems. Auditing, stats, and access logging come free with every endpoint via middleware, not as an afterthought. The plugin architecture lets you add private extensions without forking the repo.
It's being archived by Lyft — no new features, no PRs accepted. That's a hard stop for adoption; you'd be taking on a dead dependency the moment you build on it. The Protobuf-everywhere approach means a non-trivial toolchain (buf, protoc, generated Go and TypeScript) just to add a workflow. The chaos experimentation module is deeply Lyft-specific (serverexperimentation, redisexperimentation) and won't translate to most environments without significant rework. With only 1774 stars and 124 forks, the community never reached the size needed to sustain this without Lyft behind it.