twz123/kine

Kine lets you run Kubernetes using a database you probably already have, like PostgreSQL, MySQL, or SQLite, instead of the default database Kubernetes normally requires. Normally, Kubernetes depends on a specialized database called etcd to store its internal state and coordinate the system. Kine acts as a translator in the middle: Kubernetes thinks it is talking to etcd, but Kine is actually routing those requests to a more familiar database. This matters because etcd can be a pain to operate, especially for smaller teams or single-node setups. If you'd rather just point Kubernetes at a Postgres instance or a local SQLite file, Kine makes that possible. It was originally built for K3s, a lightweight Kubernetes distribution, but it can run standalone so any Kubernetes cluster can use it. The project works by implementing just enough of the etcd API to keep Kubernetes happy. It translates the operations Kubernetes expects, create, update, delete, into equivalent calls against whichever database you've configured. The README is upfront that this is a subset of etcd's full API, so it's not a general-purpose etcd replacement, it's specifically tuned for what Kubernetes needs. Who would use this? Someone running Kubernetes in a constrained environment, a laptop, a small server, or a setup where you already have a Postgres or MySQL instance and don't want to stand up a separate etcd cluster just for Kubernetes. It's a practical tool for people who want Kubernetes but with simpler infrastructure underneath. The README is sparse on operational details, so you'd likely need to check the provided example and flow diagram to understand setup and limitations before committing to it.