Kubernetes Troubleshooting Cheat Sheet

kubectl get pods -n payments -o custom-columns=NAME:.metadata.name,PHASE:.status.phase,NODE:.spec.nodeName,IP:.status.podIP

# Render table output with selected fields.

kubectl get deployment web -n payments -o json

# Dump a resource as JSON.

kubectl get deployment web -n payments -o yaml

# Dump a resource as YAML.

kubectl get pods -n payments --sort-by=.status.startTime

# Sort list output by a JSONPath-like field.

kubectl get pods -A --field-selector=status.phase!=Running

# Filter server-side using field selectors.

kubectl get pods -n payments -l 'app=web,component=api'

# Filter resources using labels.

kubectl get pods -n payments --no-headers

# Suppress table headers for scripting.

kubectl get pods -n payments -o name

# Return resource identifiers only.

kubectl debug node/worker-02 -it --image=busybox:1.36

# Launch a debugging pod on a node.

kubectl debug web-abc123 -n payments --copy-to=web-debug --container=web -- sh

# Create a debug copy of a pod.

kubectl debug -it web-abc123 -n payments --image=nicolaka/netshoot --target=web

# Inject an ephemeral debug container into a running pod.

kubectl auth can-i create deployments -n payments

# Ask the API server whether an action is allowed.

kubectl auth can-i get secrets -n payments --as=system:serviceaccount:payments:web

# Test RBAC from another subject's perspective.

kubectl get pod web-abc123 -n payments -o jsonpath='{range .status.conditions[*]}{.type}{"="}{.status}{"
"}{end}'

# Print pod condition states.

kubectl describe pods -n payments -l app=worker

# Inspect all matching pods for failure clues.

kubectl get pods -A --field-selector=status.phase=Pending

# Find pods waiting to schedule or start.

kubectl describe pod pending-pod -n payments

# Look for failed scheduling and image pull events.

kubectl get nodes -o custom-columns=NAME:.metadata.name,READY:.status.conditions[?(@.type=="Ready")].status,MEMORYPRESSURE:.status.conditions[?(@.type=="MemoryPressure")].status,DISKPRESSURE:.status.conditions[?(@.type=="DiskPressure")].status

# Check high-level node health indicators.

kubectl rollout pause deployment/web -n payments

# Temporarily pause a deployment rollout.

kubectl rollout resume deployment/web -n payments

# Resume a paused deployment rollout.

kubectl get pdb -A

# Show PodDisruptionBudgets.

kubectl describe pdb web -n payments

# Inspect disruption constraints that may block eviction.

kubectl get pods -A --field-selector=status.phase=Failed | grep Evicted | awk '{print $1, $2}'

# Pattern to find evicted pods for cleanup.

kubectl delete jobs -n ops --field-selector=status.successful=1

# Clean up successful Jobs when appropriate.

kubectl get namespace stuck-ns -o jsonpath='{.spec.finalizers}'

# See whether finalizers are blocking deletion.

kubectl get --raw='/healthz?verbose'

# Read API server health information.

kubectl get --raw='/readyz?verbose'

# Read API server readiness information.

kubectl get componentstatuses

# Legacy command seen on older clusters; modern clusters rely on health endpoints and metrics.