Kine
by k3s-io
Kine: Replace etcd with SQL Databases in Kubernetes
Drop-in etcd replacement that lets Kubernetes clusters use MySQL, PostgreSQL, or SQLite as their datastore backend.
- 2,250+ GitHub stars
- Built with Go
- Full etcd API v3 compatibility with transparent SQL translation layer
- Apache License 2.0 license
About This Project
Kine is a shim that translates etcd's API calls into SQL database operations, enabling Kubernetes to run on familiar relational databases instead of etcd. This breakthrough simplifies infrastructure requirements and leverages existing database expertise for managing Kubernetes state.
Traditional Kubernetes deployments require etcd for cluster state management, adding operational complexity and specialized knowledge requirements. Kine eliminates this dependency by providing a transparent translation layer that works with MySQL, PostgreSQL, and SQLite, allowing teams to use battle-tested databases they already know and trust.
Originally developed for K3s (lightweight Kubernetes), Kine has proven particularly valuable for edge deployments, development environments, and scenarios where minimizing infrastructure footprint matters. It maintains full API compatibility while dramatically reducing the operational burden of running Kubernetes clusters.
The project's SQL-based approach opens new possibilities for backup strategies, replication setups, and high-availability configurations using standard database tooling rather than etcd-specific solutions.
Key Features
- Full etcd API v3 compatibility with transparent SQL translation layer
- Support for MySQL, PostgreSQL, and SQLite database backends
- Zero application changes required - drop-in replacement for etcd
- Significantly reduced resource footprint compared to etcd clusters
- Leverage standard SQL database backup, replication, and HA tools
- Seamless integration with K3s and other Kubernetes distributions
- Single-node simplicity with SQLite or multi-node scalability with SQL databases
How You Can Use It
Running lightweight Kubernetes clusters with SQLite for edge computing and IoT deployments
Leveraging existing PostgreSQL or MySQL infrastructure for Kubernetes state management
Simplifying Kubernetes development environments without etcd overhead
Building cost-effective K8s clusters using managed database services from cloud providers
Enabling Kubernetes on resource-constrained systems where etcd is too heavy
Creating portable single-binary Kubernetes distributions for testing and demos
Who Is This For?
Platform engineers, DevOps teams, and developers working with K3s, edge computing, or lightweight Kubernetes deployments who want to reduce infrastructure complexity