Day 37 - Application Troubleshooting in Kubernetes

 Day 37 - Application Troubleshooting in Kubernetes

#Day-37
Today I worked on Application Troubleshooting in Kubernetes — deep-diving into how microservices communicate, how services expose workloads, and how to resolve real-world connectivity issues inside a cluster!

Key Learnings & Fixes:
1️⃣ Deployed the Example Voting App, which consists of 5 microservices — frontend, Redis, .NET worker, PostgreSQL, and results UI.
2️⃣ Verified pod placement using kubectl get po -o wide and traced traffic flow across worker nodes to identify NodePort accessibility issues.
3️⃣ Identified missing service endpoints caused by label mismatches — fixed by aligning pod labels with service selectors.
4️⃣ Debugged Result Service connectivity — the pod listened on port 80 but the service was targeting 8080, so I patched the service using kubectl edit svc result.
5️⃣ Diagnosed vote-submission failures originating from NetworkPolicy restrictions — the policy allowed only app=frontend, while our voting pods used labels like app=vote.
6️⃣ Updated the NetworkPolicy to include the correct labels, restoring connectivity between the voting app and Redis.
7️⃣ Validated everything end-to-end: NodePort access ✔️, service endpoints ✔️, Redis communication ✔️, vote submission ✔️, result page ✔️

Pro Tips Learned Today:
1️⃣ Always validate labels and selectors — most service issues come from mismatches.
2️⃣ If a service shows no endpoints, check pod labels first before diving deeper.
3️⃣ For NodePort issues, confirm node public IP + correct worker placement.
4️⃣ NetworkPolicies can silently block communication — always verify your ingress rules.
5️⃣ Debugging in Kubernetes isn’t guessing — it’s following the network path step-by-step.

ReadMore: https://lnkd.in/dFn4XKHH