How Can I Locate the Root Cause If a Workload Is Abnormal?

Fault Locating

1. Check whether the workload pod is running properly.
   1. Log in to the CCE console.
   2. Click the cluster name to access the cluster console. In the navigation pane, choose Workloads.
   3. In the upper left corner of the page, select the namespace, locate the target workload, and view its status.
      • If the workload is not ready, you can view pod events to determine the cause. For details, see Viewing Pod Events. You can find the solution to the issue based on the events by referring to Common Pod Issues.
      • If the workload is processing, wait patiently.
      • If the workload is running, but it is not accessible, check whether intra-cluster access is normal.

2. Check whether access within the cluster is normal.

   Log in to the CCE console or use kubectl to obtain the pod IP address. Then, log in to the node or the pod and run curl or use other methods to manually call the APIs and check whether the intra-cluster communication is normal.

   If {Container IP address}:{Port number} is not accessible, log in to the service container, and attempt to access 127.0.0.1:{Port number}.

3. Check whether the expected results are displayed.

   If the workload is accessible within the cluster but the expected results are not shown, check the workload configurations, such as verifying if the image tag and environment variables are correctly configured.

Common Pod Issues

Viewing Pod Events

kubectl describe pod {pod-name}

...
Events:
  Type     Reason            Age   From               Message
  ----     ------            ----  ----               -------
  Warning  FailedScheduling  49s   default-scheduler  0/2 nodes are available: 2 Insufficient cpu.
  Warning  FailedScheduling  49s   default-scheduler  0/2 nodes are available: 2 Insufficient cpu.