Kubernetes Simulator

advanced · ~12 min

The 2am failure

nightly-export has run every night for months. You're on call.

When a scheduled run fails, two jobs remain: stop the next run failing too, and make up the run that was missed. A CronJob reads its Secrets fresh each time it starts a Job — unlike a long-running pod.

Objectives

  • Find out why the export is failing
  • Put the new token in the warehouse Secret
  • Get an export to complete on the new token
  • Bonus: Backfill now with kubectl create job --from=cronjob/nightly-export

Best on a computer with a keyboard. It works on a phone, with shortcuts under the terminal.