kind
Kubernetes IN Docker - local clusters for testing Kubernetes
kind (Kubernetes in Docker) runs a full local Kubernetes cluster using Docker containers as nodes, intended for testing, CI, and local development.
Kind (short for Kubernetes IN Docker) is a tool that lets you run a full Kubernetes cluster on your own computer using Docker containers. Kubernetes is the system used to manage and run large numbers of containerized applications in production environments — typically across many servers in a data center. Kind makes it easy to spin up a local version of Kubernetes on your laptop for testing and development purposes, without needing real servers.
Each machine node in the cluster is simulated by a Docker container. This means you can create a working Kubernetes environment with a single command — kind create cluster — and delete it just as quickly. Kind also supports multi-node clusters, including high-availability setups where multiple control planes run simultaneously for resilience. It can also build Kubernetes from source code, which is useful for developers working on Kubernetes itself.
Kind is a CNCF certified conformant Kubernetes installer, meaning it passes the official Kubernetes compatibility tests. It runs on Linux, macOS, and Windows. In addition to Docker, it works with Podman and nerdctl (alternative container runtimes). It is written in Go and can be installed via Homebrew, Chocolatey, or by downloading a binary directly. The project is maintained under the Kubernetes Special Interest Group for Testing (SIG-Testing) and is supported by a community via Kubernetes Slack.
Where it fits
- Spin up a throwaway Kubernetes cluster on a laptop with one command.
- Run Kubernetes-based integration tests inside a CI job.
- Test multi-node and HA control-plane setups locally.
- Build and load custom Kubernetes images for development.