Using kubectl The Command Line Tool

kubectl is the primary tool you use to talk to a Kubernetes cluster. Every action — creating resources, checking status, debugging problems, or deleting objects — goes through kubectl. Mastering it makes you effective at every stage of the Kubernetes learning journey.

How kubectl Communicates with the Cluster

kubectl does not talk to worker nodes directly. It sends requests to the API Server on the control plane. The API Server processes the request, updates etcd, and instructs the rest of the cluster. Your laptop just needs network access to the API Server's endpoint.

Your Laptop
kubectl command
     ↓
API Server (Control Plane)
     ↓
etcd / Scheduler / Controller
     ↓
Worker Nodes run your workloads

The kubeconfig File

kubectl reads a file called kubeconfig to know which cluster to connect to and what credentials to use. This file lives at ~/.kube/config by default. When you run minikube start, Minikube writes this file automatically. When you download credentials for a cloud cluster (GKE, EKS, AKS), the provider updates this file.

To see all clusters your kubectl knows about:

kubectl config get-contexts

To switch to a different cluster:

kubectl config use-context my-production-cluster

The kubectl Command Structure

Every kubectl command follows a consistent pattern:

kubectl [action] [resource-type] [resource-name] [flags]

Examples:

kubectl get pods
kubectl get pods my-pod-name
kubectl describe deployment my-app
kubectl delete service my-service
kubectl apply -f app.yaml

The Most Important kubectl Commands

Getting Resources

Describing Resources

kubectl describe gives you detailed information about one resource — events, labels, conditions, and configuration. This is your first debugging tool when something looks wrong.

kubectl describe pod my-pod-name
kubectl describe node minikube

The events section at the bottom of the describe output often tells you exactly why a Pod failed to start.

Applying and Deleting Configurations

kubectl apply -f deployment.yaml    # Create or update from a YAML file
kubectl delete -f deployment.yaml   # Delete what the YAML file describes
kubectl delete pod my-pod           # Delete one specific Pod

Viewing Logs

Container logs help you understand what your application is doing inside the Pod:

kubectl logs my-pod-name
kubectl logs my-pod-name -f          # Follow logs in real time
kubectl logs my-pod-name --previous  # Logs from a crashed container

Running Commands Inside a Container

Sometimes you need to open a shell inside a running container to inspect files, test network connectivity, or debug an issue:

kubectl exec -it my-pod-name -- /bin/sh
kubectl exec my-pod-name -- ls /app

The -it flags make the session interactive with a terminal. This is the Kubernetes equivalent of SSH-ing into a server.

Port Forwarding

Port forwarding lets you access a Pod directly from your laptop without exposing a Service:

kubectl port-forward pod/my-pod-name 8080:80

Open http://localhost:8080 in your browser and you connect to port 80 inside the Pod. This is useful for testing and debugging without creating a Service.

Output Formats

kubectl returns plain text by default. You can request different formats for scripting or detailed inspection:

kubectl get pods -o wide         # Extra columns: IP, node name
kubectl get pods -o yaml         # Full YAML definition of the resource
kubectl get pods -o json         # JSON format for scripting
kubectl get pods -o name         # Just the resource names (pod/name)

Labels and Selectors

Labels are key-value pairs you attach to resources. Selectors let you filter resources by label. Most kubectl commands support label selectors with the -l flag.

kubectl get pods -l app=frontend
kubectl get pods -l environment=production,tier=backend

Editing Resources Live

You can open a resource directly in your terminal editor and edit it without touching any file:

kubectl edit deployment my-app

This opens the full YAML in your default editor. Save and close — Kubernetes applies the changes immediately. Use this carefully in production.

Watching Resources

The --watch flag keeps the output updated in real time, so you can monitor a deployment rolling out:

kubectl get pods --watch

Useful Shortcuts and Tips

Key Points

• kubectl talks to the API Server — it never reaches worker nodes directly.
• The kubeconfig file at ~/.kube/config stores cluster credentials and context.
• Every command follows the pattern: kubectl [action] [resource] [name] [flags].
• kubectl describe and kubectl logs are your primary debugging tools.
• Use kubectl exec -it to open a shell inside a running container for live inspection.