What Is Kubernetes?

The short answer

Kubernetes (often K8s) runs containers across a cluster of machines and keeps them running the way you described. You declare the desired state ("three copies of this image, reachable at this name, with these limits"); Kubernetes' controllers continuously compare it with the actual state and fix differences: they restart crashed containers, replace failed nodes' pods, roll out new versions and scale.

What it gives you:

• Scheduling of containers onto machines with enough CPU and memory.
• Self-healing: restarts, rescheduling, health checks.
• Service discovery and load balancing inside the cluster.
• Rolling updates and rollbacks.
• Configuration and secrets separate from images.
• Autoscaling of pods and, with cloud integrations, of nodes.

When you don't need it

A single service, a small team, or a platform-as-a-service that already runs containers for you (Cloud Run, App Runner, Container Apps, Fly.io) is often simpler. Kubernetes pays off with many services, teams and environments.

kubectl version -o json | python3 -c 'import json,sys; v=json.load(sys.stdin); print("client", v["clientVersion"]["gitVersion"], "| server", v["serverVersion"]["gitVersion"])'
kubectl get nodes -o custom-columns=NODE:.metadata.name,STATUS:.status.conditions[-1].type,KUBELET:.status.nodeInfo.kubeletVersion,RUNTIME:.status.nodeInfo.containerRuntimeVersion

client v1.37.1 | server v1.37.0
NODE                    STATUS   KUBELET   RUNTIME
rc-docs-control-plane   Ready    v1.37.0   containerd://2.3.4

Every example in this tutorial was run with kubectl against a real (throwaway) cluster when the page was built. See Kubernetes overview.