Skip to content

/glossary/kubernetes/pod

What is a Pod in Kubernetes?

The commands on this page have not been executed by our verification runner yet.updated 2026-10-20
tool
Kubernetes
versions
Kubernetes 1.x
os
any

A Pod is the smallest unit Kubernetes schedules and runs. It wraps one or more containers that always run on the same node and share a network namespace (one IP address, so they talk over localhost) and the Pod's volumes. Most Pods have a single application container, sometimes with helper (sidecar or init) containers.

A minimal Pod

yaml
apiVersion: v1
kind: Pod
metadata:
  name: web
  labels:
    app: web
spec:
  containers:
    - name: nginx
      image: nginx:1.29
      ports:
        - containerPort: 80

You rarely create Pods directly: a Deployment (through a ReplicaSet), StatefulSet, DaemonSet or Job creates them from a Pod template and replaces them when they fail or when a node goes away. Pods are disposable and get a new name and IP when recreated, which is why traffic goes through a Service.

Pod phases

PhaseMeaning
Pendingaccepted, but not all containers are running yet (scheduling, image pull)
Runningbound to a node, at least one container is running
Succeededall containers exited with status 0 and will not restart
Failedall containers ended and at least one failed
Unknownthe node cannot be reached

CrashLoopBackOff and ImagePullBackOff are not phases but container states (reasons) shown by kubectl get pods; kubectl describe pod NAME shows the events behind them.

Sources