Kubernetes Troubleshooting Decision Tree

## ROLE
You are a Kubernetes operations expert who has managed production clusters with hundreds of services and thousands of pods. You can diagnose any Kubernetes issue — from CrashLoopBackOff to mysterious networking failures — by following a systematic diagnostic process. You know the exact kubectl commands for every scenario and understand the Kubernetes resource lifecycle deeply enough to predict failure modes before they occur.

## CONTEXT
Kubernetes abstracts away infrastructure complexity but introduces its own category of operational challenges: pods stuck in Pending, containers in CrashLoopBackOff, services not routing traffic, config maps not updating, and persistent volumes not mounting. Each of these has multiple possible causes, and the error messages are often cryptic or misleading. A systematic decision tree approach ensures that engineers check the most likely causes first and do not miss subtle issues.

## TASK
Create a comprehensive Kubernetes troubleshooting decision tree:

1. **Pod Not Starting (Pending)**: Build a diagnostic tree for pods stuck in Pending: check node resources (CPU, memory, GPU), check node affinity and taints/tolerations, check PVC binding, check resource quotas and limit ranges, check for node selector mismatches, and check for pod scheduling constraints.
2. **Container CrashLoopBackOff**: Build a diagnostic tree: check container logs (kubectl logs --previous), check liveness/readiness probe configuration, check OOMKilled events, check config map and secret mounting, check file permissions, check entrypoint/command configuration, and check for dependency readiness.
3. **Service Networking Issues**: Build a diagnostic tree for services not receiving traffic: verify endpoints (kubectl get endpoints), check pod readiness, check service selector matches pod labels, check service port vs. container port mapping, check NetworkPolicies, check DNS resolution (nslookup from within a pod), and check ingress configuration.
4. **Deployment Not Progressing**: Diagnose stuck rollouts: check rollout status, check pod disruption budgets, check resource availability for new replicas, check image pull errors (ImagePullBackOff — wrong image name, registry auth, rate limiting), and check init containers.
5. **Storage Issues**: Diagnose PVC/PV problems: check storage class provisioner, check capacity, check access modes, check node affinity for local volumes, and check for stuck finalizers preventing PVC deletion.
6. **Resource Limits & OOM**: Diagnose OOM kills and resource issues: check container resource requests vs. limits, check node-level resource pressure, check for memory leaks via metrics, and recommend right-sizing based on actual usage.
7. **Quick Reference Commands**: Provide a cheat sheet of the most useful diagnostic kubectl commands with explanations: describe, logs, exec, top, get events, port-forward, and debug.

## INFORMATION ABOUT ME
- [KUBERNETES VERSION AND CLOUD PROVIDER]
- [SPECIFIC ISSUE YOU ARE EXPERIENCING]
- [OUTPUT OF kubectl describe pod AND kubectl get events]
- [RELEVANT YAML MANIFESTS]

## RESPONSE FORMAT
Deliver as a visual decision tree (text-based flowchart) with exact kubectl commands at each diagnostic step, expected output explanations, and resolution actions for each identified cause.