kubectl-delete

Delete a specific pod

$ kubectl delete [[po|pods]] [pod_name]

$ kubectl delete [[deploy|deployments]] [deployment_name]

$ kubectl delete [[no|nodes]] [node_name]

$ kubectl delete [[po|pods]] --all [[-n|--namespace]] [namespace]

$ kubectl delete [[deploy|deployments]],[[svc|services]] --all [[-n|--namespace]] [namespace]

$ kubectl delete [[no|nodes]] --all

$ kubectl delete [[-f|--filename]] [path/to/manifest.yaml]

kubectl delete (-f FILENAME | TYPE [NAME_PREFIX] | TYPE/NAME | -l label_selector | -all) [options]

-f, --filename
    Filename, directory, or URL to files identifying resources to delete.

-k, --kustomize
    Process directory with Kustomize before deletion.

-l, --selector
    Selector (label query) to filter resources.

--all
    Select all resources in the namespace of the specified type.

-R, --recursive
    Recursively delete all resources under a namespace.

--cascade
    Deletion strategy: 'background' (default), 'orphan', or 'foreground'.

--grace-period
    Period of time in seconds given to the resource to terminate gracefully.

--timeout
    Maximum time to wait for deletion completion.

-n, --namespace
    Namespace for the deletion.

--field-selector
    Selector based on resource fields.

The kubectl delete command removes one or more Kubernetes resources, such as pods, services, deployments, or namespaces, from a cluster. It supports deletion by resource name, label selector, file input, or all resources of a type. This is a core operation for managing cluster state, allowing users to clean up resources after use or correct misconfigurations.

Resources can be specified directly (e.g., kubectl delete pod mypod), via YAML/JSON files (-f file.yaml), or by labels (-l app=myapp). By default, it cascades deletion to dependent objects like ReplicaSets for Deployments. Use --cascade=orphan to detach instead. Graceful shutdown is supported with --grace-period and --timeout.

This command is essential for CI/CD pipelines, testing environments, and production maintenance, ensuring resources don't linger and consume cluster capacity. Always verify with kubectl get before deletion to avoid accidents. Deleted resources are permanently gone unless protected by finalizers or admission controllers.

Deletion is irreversible; use kubectl get first. Finalizers may block deletion. Protected resources require manual removal.

kubectl delete pod mypod
kubectl delete -f deployment.yaml
kubectl delete deployment,replicaset -l app=myapp
kubectl delete ns myns --cascade=orphan

Use --dry-run=client or server to preview deletions without applying.

Introduced with kubectl v1.0 in 2014 as part of Kubernetes. Evolved with cascade options in v1.5 (2017) for better dependency handling. Maintained by CNCF Kubernetes project.

kubectl get(1), kubectl apply(1), kubectl describe(1), kubectl(1)