Kubernetes Fundamentals
Stand up a real local Kubernetes cluster with kind, and learn the core object model — Pods, the declarative apply workflow, namespaces, and the kubectl tools you'll use every day — entirely from the command line.
Prerequisites: Docker Desktop or another Docker-compatible engine installed and running; comfort with a terminal — no prior Kubernetes experience required.
Version note: verified against kubernetes.io and kind.sigs.k8s.io docs, July 2026. Targets kind v0.32.0 (default node image ships Kubernetes 1.35/1.36) and a matching kubectl. Kubernetes itself is on the 1.36 minor line (latest patch 1.36.2) with 1.35 and 1.34 also in support under the N-2 policy — run kubectl version and kind version to confirm what you actually have before following version-specific steps elsewhere.
Kubernetes is a control plane that continuously reconciles the state you declare ("I want 3 replicas of this container") against the state that actually exists, restarting, rescheduling, and rerouting traffic as needed. You can't get a real feel for that reconciliation loop by reading about it — you need a cluster you can poke at, break, and rebuild in seconds.
kind ("Kubernetes IN Docker") runs full Kubernetes nodes as Docker containers on your machine. It was built by the Kubernetes project itself to test Kubernetes, so it tracks upstream releases closely and is a trustworthy stand-in for a real cluster. In this project you'll create a kind cluster, then use kubectl to create Pods imperatively and declaratively, inspect objects, and clean up. Everything here transfers directly to a managed cluster (EKS, GKE, AKS) later — only the cluster-creation step changes.
Test what you just learned
Self-testing is one of the best ways to retain new skills. Unlock project quizzes to check your understanding.
Log in to unlock0 / 7 complete