fixes core#2687: regression on group rebuild scheduled job

[MegaphoneJon]

Overview

Cron fails if you have "Rebuild Smart Group Cache" enabled in the default configuration. Detailed replication steps on the ticket: https://lab.civicrm.org/dev/core/-/issues/2687

Before

Cron fails (specifically, job.rebuild_group fails).

After

Cron succeeds.

Comments

This is a side effect of a refactoring in #20373.

[civibot]

(Standard links)

• If this is your first pull-request for CiviCRM, please browse CONTRIBUTING.md for information about the development and testing processes.
• If you are reviewing this pull-request, you may wish to consult the test sites and the Review Standards (long template, short template).

[eileenmcnaughton]

@MegaphoneJon can you put this against the rc ?

[MegaphoneJon]

done...once I remembered the rc is now 5.40

[MegaphoneJon]

It's also reaching my attention that this configuration may have just been silently failing in the past. I'm a bit worried about what's going to happen when all of a sudden various sites with inefficient smart group queries start running them...

[eileenmcnaughton]

@MegaphoneJon hmm - that's a good point - this job is one that should only be enabled with care on sites for which it is suitable. I recently realised people thought it was recommended to use this job - but I think the docs are clearer on that now.

How sure are you that it was silently failing?

[MegaphoneJon]

Not 100% sure - but after deploying this fix to production on two sites, their cron jobs started failing. When I ran my smart group profiling script, I found smart groups that couldn't possibly have finished in a reasonable timeframe.

[eileenmcnaughton]

@andrew-cormick-dockery that script ^^ would have been helpful for you in the past

[seamuslee001]

@eileenmcnaughton @MegaphoneJon so where are we at with this then?, it makes sense to me but just wondering about the concern eileen raised

[eileenmcnaughton]

We can put this in the dev digest & get more people to check

[kcristiano]

This PR also makes sense to me. I took a look at a number of sites and those that are running this job, but did not change the limit are silently failing. Based on the docs https://docs.civicrm.org/user/en/latest/initial-set-up/scheduled-jobs/#job_group_cache_flush I've started to switch sites over to disabling this job and adding group_cache_flush

As we already have documentation, I think we could merge this and add the details in the release notes. I am not sure if we need an upgrade notice as well. I could go either way on that.