Lack of visibility of job status / DB maintenance mode

[rebkwok]

On 3rd Dec we went into DB maintenance mode around 10am. A user kicked off a set of jobs later that day, while maintenance mode was still on, then realised there was an error and cancelled them. Some of these were database jobs, and remained as pending (fixing the cancel behaviour during maintenance mode is a job for the RAP team, see opensafely-core/job-runner#766). However, the user was then confused as to why some cancelled jobs were apparently still running, and pinged tech support.

It appears there is no general visibility on database maintenance mode (didn't there used to be a banner with a notification about it?). And in order to see what was going on I had to click into the job request and then the individual pending job - this was the only place I found the "waiting on database maintenance" message.

It would be useful to show more status info on the job request page.
From slack thread:

We have STATUS and STATUS_CODE in job-runner
STATUS is coarse, and is one of PENDING|RUNNING|FAILED|SUCCEEDED. It was the only status originally, which is why job-server still shows it.
STATUS_CODE is more fine grained, with 19 values.
IMO, I would like to see STATUS_CODE displayed as well as or instead of STATUS, but there may be UX considerations.

[evansd]

Just adding another case where lack of visibility of db maintenance mode has caused user confusion:
https://bennettoxford.slack.com/archives/C01D7H9LYKB/p1735648048048819

[mikerkelly]

Removing tech-support label as I don't think a user is expecting a response (the tech-support process uses this label for tracking such tickets). I leave it to Team REX processes to decide if there's anything further to do with this ticket.