I'd suggest the following note be used: Automatically closed as all accounts blocked

The current wording doesn't make it clear that:

• The accounts were not all blocked when the case was closed
• The system did this automatically without user intervention

Removing project tag WMF-NDA, not sure why that was added (it has zero influence on task visibility).

A few notes from today's meeting with @Dreamy_Jazz about this task:

• There are three blocking systems that we'll need to account for (local blocks, global blocks, locks); let's start from local blocks.
  • This applies both to reacting to the (b)lock and to checking if a user in case is (b)locked.
• The closures will be implemented as a job, configured with jobReleaseTimestamp to delay it (30-60 min should be fine).
• When reacting to global blocks and locks, we'll need to run jobs for all the wikis the current user is attached to (for example, steward locking an account on metawiki may cause a case on enwiki to have only (b)locked accounts)
  • Automatic closures on remote wikis should happen immediately, as we shouldn't expect that a steward will go and check SI on every wiki to mark case as closed (and on wikis with local CUs they shouldn't do it).

Change #1236788 had a related patch set uploaded (by Mpostoronca; author: Mpostoronca):

[mediawiki/extensions/CheckUser@master] Add auto-close job for SuggestedInvestigations cases

https://gerrit.wikimedia.org/r/1236788

Change #1236788 merged by jenkins-bot:

[mediawiki/extensions/CheckUser@master] Add auto-close job for SuggestedInvestigations cases

https://gerrit.wikimedia.org/r/1236788

Moving to "Epics in progress" to reflect that work is happening in subtasks of this task

Do we want to consider submitting the job one-off for all existing open SI cases? It feels like it could help reduce the burden of closing the older cases which already met the criteria but did so before the we had this system?

In T414979#11634530, @Dreamy_Jazz wrote:

Do we want to consider submitting the job one-off for all existing open SI cases? It feels like it could help reduce the burden of closing the older cases which already met the criteria but did so before the we had this system?

Sorry I missed this comment, do you mean write a command that we'd run one time ?

In T414979#11639741, @MPostoronca-WMF wrote:

In T414979#11634530, @Dreamy_Jazz wrote:

Do we want to consider submitting the job one-off for all existing open SI cases? It feels like it could help reduce the burden of closing the older cases which already met the criteria but did so before the we had this system?

Sorry I missed this comment, do you mean write a command that we'd run one time ?

Yes, essentially a one-use script that we run on the wikis we have suggested investigations on. Probably which just is a script that loops over every case and queues the auto-close job on it.

It could be as simple as making a script that we don't even merge into the extension and just run straight on production (to avoid the need to write tests and get the code quality good enough to merge into the extension for the long term)