Kubernetes Troubleshooting MASTERCLASS | Control Plane, Data Plane & Pod Failures
📌 Kubernetes Troubleshooting Masterclass | Control Plane, Data Plane & Pod Failures
This hands-on masterclass focuses on fixing real Kubernetes issues in the control plane, the data plane, and at the pod level. We break clusters on purpose and repair them step by step using repeatable checks and commands. You will learn how to diagnose the API server, etcd, controller manager, scheduler, kubelet, kube-proxy, Calico CNI, and CoreDNS, then tie it all together with pod lifecycle, restart and image pull behavior, and common failure patterns.
Note: This is a single cut that stitches Day 22, Day 57, and Day 58 from my CKA series. The combined GitHub notes for this masterclass are linked below; if you hear me say “Day 7” or “Day 10,” those refer to the CKA playlist sequence.
📂 GitHub Repository
All manifests and notes for this lecture:
🔗 https://github.com/CloudWithVarJosh/Y...
▶️ Watch the Full CKA 2025 Playlist
• CKA Certification Course 2025
Kubernetes Projects:
• Kubernetes Projects
Join the Channel to show your support:
/ @cloudwithvarjosh
📘 What You’ll Learn
How to read control plane health and recover static pods under /etc/kubernetes/manifests
API server diagnosis with ss, logs, health endpoints, and etcd connectivity checks
etcd certificate and SAN issues, data directory and volume problems, and safe recovery paths
Controller manager and scheduler troubleshooting, leader election, flags, and kubeconfigs
Data plane checks for Node NotReady, kubelet to API connectivity, and CRI sanity with crictl
Calico CNI basics, why pods get stuck in ContainerCreating, and fixing IPAM and MTU issues
kube-proxy modes, empty endpoints, and why a ClusterIP curl can fail
CoreDNS logs and config, upstream resolvers, and common DNS lookup failures
Pod termination signals, lifecycle states, restart policies, and image pull policies
Common errors like CrashLoopBackOff, ImagePullBackOff, and how to verify with kubectl describe, events, and logs
⏰ Timestamps
00:00:00 Introduction
00:03:17 Control Plane Troubleshooting Prerequisites
00:06:05 Control Plane and Static Pods Quick Tour
00:17:15 Scenario 1: API Server, diagnose with ss (socket stats) and logs
00:31:52 Scenario 2: etcd, certificates and SANs
00:39:37 Scenario 3: etcd, data directory and volume issues
00:46:12 Scenario 4: Controller Manager recovery
00:51:38 Scenario 5: Scheduler unreachable or misconfigured
00:56:58 Scenario 6: Renewing expired certificates with kubeadm
01:06:31 Data Plane Troubleshooting Prerequisites
01:09:30 Data Plane Quick Tour
01:15:05 Cluster Overview
01:16:36 Scenario 1: Node NotReady, new pods avoid the node
01:22:13 Scenario 2: Node NotReady, exec and logs stall
01:30:08 Scenario 3: Node NotReady, kubelet connection errors
01:34:56 Scenario 4: Pods stuck ContainerCreating
01:39:08 Scenario 5: ClusterIP request fails
01:44:33 Scenario 6: DNS lookup fails
01:52:21 Pod Termination, deletion and termination signals
01:56:41 Pod Restart Policy, Always, OnFailure, Never
02:05:08 Pod Image Pull Policy, Always, IfNotPresent, Never
02:14:16 Pod Lifecycle, Pending, Running, Succeeded, Failed, Unknown
02:22:54 Demo, observing short lived and longer pod lifecycle
02:31:25 Outro
🔗 Stay Connected
👉 LinkedIn: / varun-joshi-2b516752
👉 GitHub: https://github.com/CloudWithVarJosh
💬 Got questions? Drop them in the comments — I reply to every single one.
👍 Like, share, and subscribe to support the channel.
🔖 Hashtags
#Kubernetes #CKA2025 #Troubleshooting #ControlPlane #DataPlane #Pods #kubeadm #containerd #Calico #CoreDNS #kubeproxy #kubelet #etcd #CrashLoopBackOff #ImagePullBackOff #K8sHandsOn #DevOps #CloudWithVarJosh