Project obsolete? .spec.ttlSecondsAfterFinished can do the similar thing

[lzfxxx]

What's the difference of using this CronJob from adding the .spec.ttlSecondsAfterFinished in Job config?

[hjacobs]

This project was initially created when the mentioned feature did not exist yet, I'll have to check if kube-job-cleaner is now completely obsolete (we also still run it in our Zalando clusters).

[hjacobs]

Apparently we can also get rid of kube-job-cleaner within Zalando with Kubernetes 1.12 and the following PR: zalando-incubator/kubernetes-on-aws#1668

[nielsole]

while we are not yet on 1.12 we also turned kube-job-cleaner off, as the really nasty cronjobs without backoffs with restart:always and the likes have been taken care off with default values in recent releases.
Thank you very much, @hjacobs it was a pleasure :)

[hjacobs]

@nielsole thanks! Just for reference, I think kube-job-cleaner only really becomes obsolete with the TTL setting in 1.12 to have a comparable feature, here the official KEP with all info: https://github.com/kubernetes/enhancements/blob/master/keps/sig-apps/0026-ttl-after-finish.md

[StevenACoffman]

In our continuous delivery environment, we still find it useful to run a lightly modified version of this code to clean up deployments past a certain age. Thanks! kube-deploy-cleaner

[hjacobs]

@StevenACoffman check out my project kube-janitor to clean up arbitrary objects after some time (either with TTL annotation or custom rule for any kind of object).