You don't have a forgetting problem.
You have a remembering problem.
Introducing DailyBrief — an AI memory assistant that works through WhatsApp.
Send messages naturally. It remembers what matters and brings it back when you need it.
Theme : One Dark Pro
Course : udemy.com/course/docker-and-…
4
55
Two days of hands-on learning reinforced one thing:
Knowing Kubernetes commands isn't enough.
Understanding what kubelet, the scheduler, controllers, and container runtime are doing behind the scenes makes debugging much easier.
2
10
Deliberately introduced failures.
ImagePullBackOff — Kubernetes couldn't pull the specified image.
CrashLoopBackOff — the container started but repeatedly crashed.
Understanding the difference between a failed image pull and a crashing application was particularly useful.
1
2
12
Moved into troubleshooting.
Practiced using:
kubectl logs
kubectl logs --previous
kubectl describe pod
kubectl get events
kubectl exec
Also inspected the container from inside the Pod.
1
2
10
One important observation:
A container can be Running but its Pod can still be NotReady.
Readiness failures don't restart containers. Liveness failures do.
Startup probes also prevent liveness and readiness checks from interfering with slow application initialization.
1
2
10
Explored Kubernetes health probes:
Startup: Has the application finished initializing?
Liveness: Is the application still healthy?
Readiness: Can it receive traffic?
Experimented with HTTP, TCP, and exec-based checks.
1
2
12
Experimented with container restart policies:
Always
OnFailure
Never
Observed how Kubernetes handles successful and failed container exits.
Also learned that restarting a container inside a Pod is different from replacing the Pod itself.
1
2
13
Deployed my first Nginx Pod.
Used kubectl apply, inspected it with kubectl describe, and accessed it through port-forwarding.
Then deleted it.
The Pod was gone.
That demonstrated an important concept: a standalone Pod has no controller to recreate it.
1
2
15
Set up a local Kubernetes cluster using kind.
Inspected nodes system Pods and control-plane components.
Kubernetes control-plane components can run as Static Pods managed directly by kubelet.
This helps bootstrap the cluster without depending on the API Server.
1
2
20
Started with Kubernetes architecture.
Explored the Control Plane and Worker Nodes, including:
API Server
etcd
Scheduler
Controller Manager
kubelet
kube-proxy
Container Runtime
Also understood how these components coordinate to maintain the cluster's desired state.
1
2
26
Learning Kubernetes by building, not just watching tutorials.
Over the last 2 days, I moved from understanding Kubernetes architecture to running experiments on a local kind cluster.
@Hiteshdotcom @techie_piyush @nirudhuuu
Here's what I explored and what finally clicked. 👇
4
21
421
@Hiteshdotcom sir and @techie_piyush sir.
2
30
Next, I want to go deeper into Kubernetes.
Not just learning kubectl commands or copying YAML files.
I want to understand what happens internally when applications are deployed, scaled, updated, or when something fails.
More hands-on learning ahead.
2
13
The role-play exercise was a good reality check.
I could explain individual concepts, but struggled to clearly describe the complete request flow and how Kubernetes components interact.
Identified an important gap to work on through more practical scenarios.
1
2
13
One important takeaway:
Deployments manage ReplicaSets.
ReplicaSets maintain the desired number of Pods.
Services provide stable connectivity to Pods through labels and selectors.
Understanding these relationships matters more than memorizing commands.
1
1
10
Explored Kubernetes Services and networking:
→ ClusterIP for internal communication
→ NodePort for external access
→ LoadBalancer for external traffic
→ NGINX and load balancing
Also deployed sample applications and services.
1
1
18
Got hands-on with two ways of managing Kubernetes resources:
→ Imperative: using kubectl commands
→ Declarative: defining desired state through YAML manifests
Also learned how Kubernetes continuously reconciles actual state with desired state.
1
2
20




