fix: Make sure that StatusUpdateTrigger and InfoUpdateTrigger keep running on rolling updates

[cdprete]

Fixes #5450 (comment).

Better titles are welcome :D

If approved and merged, it would be great to have this released as soon as possible.
I'm currently running in DEV with a patched local version (the one in this PR) and in QA&PROD I had to destroy and recreate the deployment to re-trigger the fetching of the instances, therefore those envs are good until one Pod dies or if I've to redeploy something.

[cdprete]

HI @SteKoe @ulischulte @hzpz,
can we have this reviewed as soon as possible? Thanks in advance.

[SteKoe]

Three risk findings from the review — all addressed with concrete suggestions inline.

[cdprete]

Three risk findings from the review — all addressed with concrete suggestions inline.

Thanks. I added my comments to them.

Btw, a better version would be to have IntervalTrigger have an Hazelcast IMap as well, so that not even the repull of the statuses and infos would be necessary, but that would require way more refactoring and I'm not certain (to try out somehow) it can be introduced in a non-breaking way.

@SteKoe What's your opinion on this?

[cdprete]

@SteKoe unless you want to consider #5663 (comment), all the comments have been addressed.