fix: workaround for long certificate lifetimes

[dervoeti]

Description

Workaround for #362

Tested by temporary setting the maximum delay to 1 minute (instead of 6 months) and requesting a cert with a couple minutes lifetime. It rescheduled a few times and then restarted the Pod with a new cert, so the Pod lived for multiple minutes even though the maximum delay was 1 minute, see the recheck_delay in these logs:

2025-06-24T13:47:21.234336Z  INFO reconciling object{object.ref=Pod.v1./example-secret-consumer-0.default object.reason=reconciler requested retry}: stackable_commons_operator::restart_controller::pod: Pod still valid, rescheduling check pod.expires_at=Some(2025-06-24T13:55:11.594779133+00:00) recheck_delay=60s
2025-06-24T13:48:21.237989Z  INFO reconciling object{object.ref=Pod.v1./example-secret-consumer-0.default object.reason=reconciler requested retry}: stackable_commons_operator::restart_controller::pod: Pod still valid, rescheduling check pod.expires_at=Some(2025-06-24T13:55:11.594779133+00:00) recheck_delay=60s
2025-06-24T13:49:21.240752Z  INFO reconciling object{object.ref=Pod.v1./example-secret-consumer-0.default object.reason=reconciler requested retry}: stackable_commons_operator::restart_controller::pod: Pod still valid, rescheduling check pod.expires_at=Some(2025-06-24T13:55:11.594779133+00:00) recheck_delay=60s
2025-06-24T13:50:21.244534Z  INFO reconciling object{object.ref=Pod.v1./example-secret-consumer-0.default object.reason=reconciler requested retry}: stackable_commons_operator::restart_controller::pod: Pod still valid, rescheduling check pod.expires_at=Some(2025-06-24T13:55:11.594779133+00:00) recheck_delay=60s
2025-06-24T13:51:21.247748Z  INFO reconciling object{object.ref=Pod.v1./example-secret-consumer-0.default object.reason=reconciler requested retry}: stackable_commons_operator::restart_controller::pod: Pod still valid, rescheduling check pod.expires_at=Some(2025-06-24T13:55:11.594779133+00:00) recheck_delay=60s
2025-06-24T13:52:21.250501Z  INFO reconciling object{object.ref=Pod.v1./example-secret-consumer-0.default object.reason=reconciler requested retry}: stackable_commons_operator::restart_controller::pod: Pod still valid, rescheduling check pod.expires_at=Some(2025-06-24T13:55:11.594779133+00:00) recheck_delay=60s
2025-06-24T13:53:21.254136Z  INFO reconciling object{object.ref=Pod.v1./example-secret-consumer-0.default object.reason=reconciler requested retry}: stackable_commons_operator::restart_controller::pod: Pod still valid, rescheduling check pod.expires_at=Some(2025-06-24T13:55:11.594779133+00:00) recheck_delay=60s
2025-06-24T13:54:21.257055Z  INFO reconciling object{object.ref=Pod.v1./example-secret-consumer-0.default object.reason=reconciler requested retry}: stackable_commons_operator::restart_controller::pod: Pod still valid, rescheduling check pod.expires_at=Some(2025-06-24T13:55:11.594779133+00:00) recheck_delay=50.33773894s

Which confirms certificate lifetimes greater than the new maximum delay of 6 months still work as expected.

Definition of Done Checklist

• Not all of these items are applicable to all PRs, the author should update this template to only leave the boxes in that are relevant
• Please make sure all these things are done and tick the boxes

Author

• Changes are OpenShift compatible
• CRD changes approved
• CRD documentation for all fields, following the style guide.
• Helm chart can be installed and deployed operator works
• Integration tests passed (for non trivial changes)
• Changes need to be "offline" compatible
• Links to generated (nightly) docs added
• Release note snippet added

Reviewer

• Code contains useful comments
• Code contains useful logging statements
• (Integration-)Test cases added
• Documentation added or updated. Follows the style guide.
• Changelog updated
• Cargo.toml only contains references to git tags (not specific commits or branches)

Acceptance

• Feature Tracker has been updated
• Proper release label has been added
• Links to generated (nightly) docs added
• Release note snippet added
• Add type/deprecation label & add to the deprecation schedule
• Add type/experimental label & add to the experimental features tracker

[lfrancke]

The original error was when someone requested a cert with 3 years lifetime which leads to a crash. Have you tested this as well?

[dervoeti]

The original error was when someone requested a cert with 3 years lifetime which leads to a crash. Have you tested this as well?

Yes, commons-operator crashlooped before, with this fix it is stable.

My upstream PR just got approved, so we can also decide not to merge this and wait for a new kube-rs release.
This fix is tested though, we could merge it before the next SDP release if a new kube-rs version is not released until then.

[lfrancke]

Yes, please still merge and create a followup ticket and send it to me.

[razvan]

in case you still want to merge this, it lgtm.

[dervoeti]

Follow up issue: #366