What We’ve Covered So Far

In the last few blogs, we’ve taken major strides:

Part 1: Understood what Kubernetes is and why it exists using real-world analogies.
Part 2: Explored the Kubernetes Architecture, both control plane and worker nodes.
Part 3: Set up Minikube on your local system and deployed your first pod using YAML.

Now that you have a working cluster and you’ve seen your first pod running, it’s time to go deeper.

In this blog, we’ll learn:

How a Pod’s lifecycle works from creation to deletion
How to use kubectl to inspect, debug, and manage pods
Common pod states and what they mean
Real-time debugging with logs, exec, and describe

What is a Pod, Again?

A Pod is the smallest deployable unit in Kubernetes.

It wraps one or more containers with:

Shared network namespace
Shared storage volumes
One IP address

In most use cases, a Pod = One Container.

Pod Lifecycle Phases

When you apply a manifest or create a Pod, it goes through a defined series of phases:

1. Pending

Pod is accepted by the cluster but not yet running.
Reasons:
- Scheduler hasn’t assigned it to a node
- Container images still downloading

2. Running

Pod is scheduled and containers are executing.
Containers have passed readiness checks (if defined).

3. Succeeded

All containers terminated successfully (exit code 0).
Seen in Jobs, not typical in long-running pods.

4. Failed

One or more containers exited with non-zero code.

5. CrashLoopBackOff

Container failed, restarted, failed again in a loop.
Common cause: missing ENV vars, database connection failure, etc.

Using kubectl to Explore Pod Lifecycle

View All Pods

kubectl get pods

With more details:

kubectl get pods -o wide

Describe a Pod

kubectl describe pod <pod-name>

Shows node info, events, reasons for failure/success, environment, etc.

Check Pod Logs

kubectl logs <pod-name>

For multi-container pods:

kubectl logs <pod-name> -c <container-name>

Exec into a Running Pod

kubectl exec -it <pod-name> -- /bin/bash

This lets you debug live, check file system, curl endpoints, run tests, etc.

Restart a Pod (delete and recreate)

kubectl delete pod <pod-name>

If created by a Deployment, it will automatically spin back up

Real-Life Example: Simulate a CrashLoop

Let’s run a broken pod that always crashes:

crashy-pod.yaml

apiVersion: v1
kind: Pod
metadata:
  name: crash-loop
spec:
  containers:
  - name: fail-container
    image: busybox
    command: ["/bin/sh", "-c", "exit 1"]

Apply:

kubectl apply -f crashy-pod.yaml

Check:

kubectl get pods

Output:

crash-loop   0/1     CrashLoopBackOff   3 (5s ago)   1m

Inspect:

kubectl describe pod crash-loop
kubectl logs crash-loop

Delete:

kubectl delete pod crash-loop

What’s Next?

In the next blog, we’ll cover:

Deploying your app using a Deployment and ReplicaSet
Scaling up/down your app
Updating your app without downtime using rolling updates

Kubernetes From Zero to Hero – Part 4: Understanding Pod Lifecycle and Deep-Diving Into Kubectl