kubectl apply -k: Usage, Options & Common CI Errors
Render a kustomize overlay and apply it in a single command.
kubectl apply -k builds a kustomize directory (base plus overlay) and applies the result, the standard way to deploy environment-specific config without templating. In CI you point -k at the right overlay per environment.
What it does
kubectl apply -k DIR runs the embedded kustomize build on DIR's kustomization.yaml - merging the base, applying patches, image tags, name prefixes, and common labels - then applies the rendered manifests. It is equivalent to kustomize build DIR | kubectl apply -f -.
Common usage
kubectl apply -k overlays/prod
kubectl apply -k overlays/staging --server-side
kustomize build overlays/prod | kubectl diff -f - # preview the change
kubectl kustomize overlays/prod # render without applyingCommon errors in CI
"error: accumulating resources ... no such file or directory" means a path in resources: is wrong or the referenced base moved - kustomize paths are relative to the kustomization.yaml. "must build at directory: not a valid directory" means you pointed -k at a file, not the folder containing kustomization.yaml. A subtle one: kubectl's built-in kustomize can lag the standalone kustomize binary, so a newer field (e.g. a transformer) may parse with kustomize build but error under apply -k - pin the kubectl version or pipe the standalone binary instead.
Using this in CI
A runner has no kubeconfig, no cached context, and no interactive auth. Every kubectl invocation in CI needs the context supplied explicitly, and most confusing CI failures here are the command running against the wrong cluster or no cluster at all.
# never rely on the ambient context on a runner
kubectl --context "$KUBE_CONTEXT" -n "$NAMESPACE" get pods
# confirm what you are actually connected to before mutating anything
kubectl config current-context
kubectl cluster-info
# fail fast instead of hanging on an unreachable API server
kubectl --request-timeout=30s get nodes