Troubleshooting - Application Failure

  • Database pod hosts a database application and serves the web servers via an application.

    • The web server is hosted on a web pod and serves users.

  • Its good to write down and map out how the application is configured before starting.

    • For example –> DB –> DB-Service –> WEB –> WEB-Service –> User

      • For troubleshooting, check every object and link in this map to find the root cause of the issue.

  • For example, we have users whom are having difficulties in accessing the application.

    • If it is a web application, check that the IP port is available by the user using curl:
curl http://web-service-ip:node-port

* Next, check the web service - has it discovered endpoints for the `WEB` pod?

    * You can use `kubectl describe serice web-service`.
    
* Compare the `Selector` configured on the service (output from the above `kubectl describe` command) to the ones on the pod.

    * Make sure the Selectors match what is in the YAML object.

  • Check the pod itself and make sure it is in a running state. The amount of restarts is good to check if the application itself is restarting or not.

  • Can also check the following:

kubectl get pod

kubectl describe pod web

kubectl logs web

  • However, one thing to keep in mind is the current version of the pod’s logs, may not match the previous fail state.

    • You can watch the logs with the -f options using kubectl logs web -f or use the --previous option to view the logs of a previous pod like so:
kubectl logs web -f --previous

  • Check the status of the DB Service.

  • Check the logs of the DB pod.

Updated: