helm-charts
helm-charts copied to clipboard
Post Upgrade / Post Install Job TTL
What would you like to be added?
The ability to configure the jobs:
- redpanda-configuration
- redpanda-post-upgrade
with a Job.spec.ttlSecondsAfterFinished (see kubectl explain job.spec
)
Why is this needed?
The standard redpanda installation comes with two jobs that help for post upgrade and configuration: i.e.:
redpanda-configuration 1/1 5s 12h
redpanda-post-upgrade 1/1 5s 12h
They leave completed pods in the namespace it's installed in for a lot longer than they need to be there. I think we could add a configuration option to allow for a TTL to be set (defaulting to the current behavior) so that the pods can optionally get cleaned up after the jobs run
Pods left unused can tie up node resources: i.e. ports or tie up cpu/mem via resources.request
JIRA Link: K8S-103