Kubernetes (K8s)

Kubernetes, often shortened to K8s, is an open-source container orchestration system. You describe the desired state of your applications in YAML manifests, and Kubernetes schedules containers onto a cluster of machines, restarts failed ones, scales them and routes traffic to them. It was open-sourced by Google in 2014 and is now a CNCF project.

How it works

A control plane stores the desired state in etcd and runs controllers that compare it with the actual state. The scheduler places pods, the smallest deployable unit of one or more containers, onto worker nodes. The kubelet agent on each node starts and watches those pods. For example, a Deployment with replicas: 3 keeps three pods running; if a node fails and one pod dies, the controller creates a replacement on a healthy node.

Why it matters in production

Kubernetes gives teams one API for deploys, scaling and self-healing across clouds and on-prem. It also adds operational work: version upgrades, node scaling, networking and observability all need deliberate design. Clusters that skip this work tend to fail during upgrades and traffic spikes.