kubectl get -o wide: Usage, Options & Common CI Errors

Add the node, pod IP, and extra columns to the default table.

kubectl get -o wide keeps the familiar table layout but adds the columns the default view omits - the node a pod landed on, its IP, and the container images. It is the quick way to see scheduling and placement at a glance.

What it does

kubectl get pods -o wide appends NODE, NOMINATED NODE, READINESS GATES, IP, and (for pods) the underlying images to the table. For nodes it adds internal/external IP, OS image, kernel, and container runtime. It is purely a richer human view - the data is the same objects.

Common usage

kubectl get pods -o wide
kubectl get pods -A -o wide                  # placement across all namespaces
kubectl get nodes -o wide                    # OS, kernel, runtime versions
kubectl get pods -o wide -l app=web | sort -k7   # group by node column

Common errors in CI

The mistake is scripting against -o wide columns: the column set and order are not a stable contract and can shift between kubectl versions, so an awk '{print $7}' that reads the NODE column silently breaks on upgrade. For scripts use -o jsonpath or -o custom-columns, which name the fields explicitly. -o wide also does not add columns for custom resources that lack additionalPrinterColumns, so a CRD may look identical to the plain table. Use it to eyeball placement (all pods on one node = an anti-affinity or topology-spread gap), then switch to jsonpath for assertions.

Related guides