Decision Tree

kubectl get pod

Scenario 1: The pod is pending

STEP: kubectl describe pod <pod-name>

DECISION POINT - are you hitting resource or quota limits

If yes, go to 1.1
If no but pod is waiting for volumes, go to 1.2
If something else go to 1.3

1.1 Hitting resource limits

STEP: Check and lower resource requests and limits

1.2 The pod waiting for Persistent Volumes

STEP: Remove PVC, they are not suppoeted in production cluster yet

1.3 Something else

STEP: kubectl describe pod -o wide

DECISION POINT - Is the Pod scheduled to a node?

If yes, go to 1.3.1
If no, go to 1.3.2

1.3.1 Pod is scheduled to a node

STEP: There might be an issue with the Kubelet. Inform ServiceOps

1.3.2 Pod is not is scheduled to a node

STEP: There might be an issue with the Scheduler. Inform ServiceOps

Scenario 2: The pod is runnig

DECISION POINT - The pod is running?

If yes, go to 2.1
If no, go to 2.2

2.1 Pod is running

STEP: kubectl logs <pod-name>

DECISION POINT - Can you see application logs?

If yes, go to 2.1.1
If no, go to 2.1.2

2.1.1 Application log available

Fix application issue causing the error log

2.1.2 Application log unavailable

STEP:kubectl get pod

DECISION POINT - Is the Pod in ImagePullBack Off?

If yes, go to 2.1.2.1
If no, go to 2.1.2.2

2.1.2.1 Application in ImagePullBack Off

STEP: Check if is the image name correct, if not fix image name STEP: Check if the image tag is correct, if not fix image tag STEP: Check if the image registry is correct, if not fix the image registry

2.1.2.2 The Pod not in ImagePullBack Off

STEP: Check if the Pod is in CrashLoopBack Off. STEP: Check application logs and fix crashing application