This turned out to be a non-bug. The observed behaviors have to do with highly variable execution times of asynchronous jobs on beta (including in particular our CacheTesterResultsJob jobs), exacerbated by the fact that many of our manual tests resulted in running UpdateImplementationJob around the same time as CacheTesterResultsJob jobs. We noticed the elapsed time between the initial creation of a job on deployment-mediawiki11 and its deployment on deployment-jobrunner04 can vary from milliseconds to 5 minutes or more. (In our manual testing we initially expected to see the results of caching available more quickly than that.)

We've determined there's no bug here. The observed behaviors resulted from:

Reopening; we've had a relapse. Caching was working reliably earlier today (based on tests manually triggered in the UI), but now it's no longer working.

@TheresNoTime - Thanks for offering to take a look! Turns out I was wrong above; in fact CacheTesterResultsJob does appear to be running. (There was some confusion about how our logging was configured on beta.) So, never mind :-). I might have one other question for you a bit later, though.

@TheresNoTime - follow-up question related to this ticket: Although UpdateImplementationsJob is now running, our extension's other job, CacheTesterResultsJob, still does not appear to be running. In your work on the config files, did you change or add something specifically related to UpdateImplementationsJob (and which maybe we need to change similarly for CacheTesterResultsJob)?

@TheresNoTime - My thanks, also!

Thanks so much, @MatthewVernon and all!

@MatthewVernon - Per my 1:1 discussion with @dr0ptp4kt earlier today, it would be good for me to have kerberos access. Apologies for not mentioning that in the original description.

Duplicates (part of) T329138

My thanks, also - @Aklapper and all!

Hi @Urbanecm @bd808 @Aklapper - I need to ask for further help from one of you. I discussed this ticket with James Forrester, who knows me personally, but is not a Phab admin. It's my understanding, from James, that any Phab admin should be able to help me, even without knowing my face, because I'm a staff member.

@Urbanecm - Yes, the MFA credentials page that I see is the same as the image you attached.

Thanks everyone for the quick response and clear explanations! I have removed my existing TOTP authentication (and will reinstate after solving the signing issue).

Superseded by T330956

Yes, it is the case. Closing.

Fixed by https://gerrit.wikimedia.org/r/c/mediawiki/extensions/WikiLambda/+/879833

I tried it on beta; seem to be working!

Issues for further consideration:

The logging is done here, in orchestrate.js: