10 Kubernetes Commands Every DevOps Engineer Uses Daily

Introduction

Kubernetes is powerful, but you don’t need to memorize 300+ kubectl commands to be effective. In real production environments in 2025, 95% of daily work is done with just 10 core commands (plus a few aliases). Whether you’re debugging a crashing pod, checking resource usage, or helping a developer access a service, these are the exact commands elite teams use hundreds of times per day.

Most clusters run securely inside a well-segmented Virtual Private Cloud (VPC) with private API endpoints.

1. kubectl get – Your Daily Dashboard

The single most used command. Always start here.

• kubectl get pods -A → see every pod in the cluster
• kubectl get pods -n production → specific namespace
• kubectl get nodes → check worker node status
• kubectl get svc,ing,deploy → quick overview

Pro tip: alias k → kubectl and use kgp for kubectl get pods

2. kubectl describe – When You Need the Full Story

kubectl describe shows events, conditions, and recent history.

• kubectl describe pod <name> → see why a pod is CrashLoopBackOff
• kubectl describe node <node-name> → check taints, capacity, events
• kubectl describe ingress <name> → verify TLS and backend rules

Events section at the bottom is pure gold during incidents.

3. kubectl logs – Follow What’s Happening Inside

Essential for debugging and monitoring.

• kubectl logs -f pod-name -c container-name → tail live logs
• kubectl logs --since=10m pod-name → last 10 minutes
• kubectl logs --previous pod-name → logs from crashed instance

Even with Loki/Grafana, engineers still reach for kubectl logs first.

Pods in private subnets still expose logs securely through the control plane or NAT Gateway-enabled agents.

4. kubectl exec – Jump Inside a Running Container

The Swiss Army knife for troubleshooting.

• kubectl exec -it pod-name -- sh
• kubectl exec -it deploy/my-app -- curl localhost:8080/health
• kubectl cp pod-name:/app/logs . → copy files

5. kubectl top – Real-Time Resource Monitoring

Instant visibility into CPU/memory usage.

• kubectl top pods -n production
• kubectl top nodes → spot overloaded workers
• Sort with --sort-by=cpu

6. kubectl port-forward – Instant Local Access

Developers love this for testing without Ingress.

• kubectl port-forward svc/my-db 5432:5432 → connect pgAdmin locally
• kubectl port-forward deploy/api 8080:80 → test API on localhost:8080

Works perfectly even when services are in private subnets – no public exposure needed.

7. kubectl scale – Instant Scaling Up/Down

For emergencies or load tests.

• kubectl scale deploy/api --replicas=20
• kubectl scale statefulset/db --replicas=0 → quick maintenance

8. kubectl apply -f – The Correct Way to Apply Changes

Never use create in production.

• kubectl apply -f manifest.yaml
• kubectl apply -k ./kustomize/overlays/prod
• GitOps tools (ArgoCD/Flux) use this under the hood

9. kubectl delete – Clean Up Fast

Sometimes you just need it gone.

• kubectl delete pod --grace-period=0 --force pod-name → kill stuck terminating pod
• kubectl delete ns stuck-namespace --grace-period=0 --force

10. kubectl config Commands – Switch Contexts Like a Pro

Real engineers manage 5–20 clusters daily.

• kubectl config get-contexts → list all clusters
• kubectl config use-context prod-us-east-1
• Alias kubectx and kubens tools save hours

Essential kubectl Cheat Sheet Table

Conclusion

Master these 10 kubectl commands and you’ll handle 99% of day-to-day Kubernetes operations with confidence. Combine them with tools like k9s, lens, or stern for even more power. The best engineers have these memorized – and so should you. Bookmark this page, add the aliases to your ~/.zshrc or ~/.bashrc, and watch your productivity soar.

Frequently Asked Questions

What is the most used kubectl command?

kubectl get pods – run hundreds of times per day by most engineers.

Should I use k9s instead of raw kubectl?

Yes for exploration, but know raw kubectl for scripts and incidents.

How do I see why a pod is pending?

kubectl describe pod <name> → look at Events section.

Can I use kubectl with multiple clusters?

Yes – manage contexts with kubectl config use-context.

Is kubectl safe on production clusters?

Yes with RBAC. Most teams grant read-only to developers and full access only to platform team.