Job's migrate pods aren't cleaned up

[copelco]

After several deploys, kubectl get pod will start to look like this:

❯ kubectl get pod | grep migrate
migrate-2fq2w               0/1     Completed   0          13h
migrate-6857f               0/1     Completed   0          12h
migrate-7kf5r               0/1     Completed   0          12h
migrate-b2745               0/1     Completed   0          12h
migrate-dt4xj               0/1     Completed   0          13h
migrate-gc6tr               0/1     Completed   0          12h
migrate-gt965               0/1     Completed   0          12h
migrate-hx4nq               0/1     Completed   0          13h
migrate-jbrjs               0/1     Completed   0          13h
migrate-kc4hq               0/1     Completed   0          12h
migrate-l6d7r               0/1     Completed   0          12h
migrate-npjjc               0/1     Completed   0          13h
migrate-v2gml               0/1     Completed   0          12h
migrate-wjqkf               0/1     Completed   0          13h

It can be convenient to see at least a few recent migrate pods to view the logs, but having them build up like this is going to get messy.

[copelco]

CronJob has successfulJobsHistoryLimit, but it's not in the Job spec as far as I can tell.

[JeremyGibson]

The CronJob and TTL Seems intriguing

[JeremyGibson]

Sorry, that was a misclick

[vkurup]

There is a setting to accomplish this, but it requires enabling a Feature flag, which seems like more customization than I'd prefer: https://kubernetes.io/docs/concepts/workloads/controllers/job/#clean-up-finished-jobs-automatically