kubectl

After this lesson, you will be able to: Use kubectl confidently: inspect cluster state, follow logs, exec into Pods, describe failures, debug issues.

kubectl is the daily-driver CLI for K8s. Master 15 commands and you can debug 90% of production issues.

Prerequisites:Core Kubernetes Concepts

Cluster inspection

Your first commands when you SSH into a cluster you've never seen.

tsx

kubectl cluster-info                  # API server endpoints
kubectl get nodes                     # cluster nodes
kubectl get nodes -o wide             # with internal IPs and OS
kubectl get namespaces                # ns ; namespaces
kubectl config current-context        # which cluster am I on
kubectl config get-contexts           # all configured clusters
kubectl config use-context my-prod    # switch

Workload inspection

These return JSON-ish lists; combine with -o for shapes.

tsx

kubectl get pods                       # default namespace
kubectl get pods -A                    # ALL namespaces
kubectl get pods -n kube-system        # specific namespace
kubectl get pods -l app=myapp          # filter by label
kubectl get pods -o wide               # show node + IP
kubectl get pods -w                    # watch (live updates)

kubectl get deployments,services,ingress -n production
kubectl get all -n production          # one-shot

Debugging individual workloads

Describe + logs + exec, the daily-driver triad.

tsx

# describe: full event history + status
kubectl describe pod myapp-abc123 -n production
# Look at the Events: section at the bottom; usually tells you why a Pod is stuck.

# Logs
kubectl logs myapp-abc123              # current logs
kubectl logs myapp-abc123 --previous   # logs from the PREVIOUS crashed container
kubectl logs -f myapp-abc123           # follow
kubectl logs -l app=myapp --tail=100   # last 100 lines across all matching pods

# Exec into a running Pod
kubectl exec -it myapp-abc123 -- bash
# Inside: same as docker exec; explore filesystem, run commands, etc.

# Run a one-shot debugging Pod
kubectl run -it --rm debug --image=nicolaka/netshoot --restart=Never -- bash

Apply, edit, delete

Modifying cluster state.

tsx

kubectl apply -f myapp.yaml            # create or update from file
kubectl apply -f k8s/                  # all YAMLs in a directory
kubectl delete -f myapp.yaml           # delete what's in the file

kubectl edit deployment myapp          # live-edit YAML (drops you in $EDITOR)
kubectl scale deployment myapp --replicas=10
kubectl rollout status deployment/myapp   # is the rollout done
kubectl rollout undo deployment/myapp     # rollback to previous version

Common mistakes only experienced engineers avoid

Forgetting the namespace. `kubectl get pods` in `default` is empty; the app's in `production`. Always check `-n`. Skipping `kubectl describe` and jumping to logs. Most failures are visible in the Events section. `kubectl edit` in production for permanent changes. Edits don't persist back to your git repo; the next CI deploy overwrites. Forgetting `--previous` on logs after a crash. The current container has fresh logs; the crashed one has the real error. Running `kubectl delete -f` on the wrong context. Always verify `current-context` before destructive operations.

Quick Check

Your Pod is in CrashLoopBackOff. What's your debug sequence?

Pick the order.

←Core Concepts: Pods, Deployments, Services

Back to Kubernetes

Deployments in Depth→