cbrain Portal-side Janitor background task

For repetitive administrative tasks, we need a control panel for launching and managing a background 'janitor' task on the portal side. Responsabilities:

Restart bourreaux after they fail, up to n times in x minutes
Trigger 'clear caches' operations for files not used in x days
Move userfiles to the archive disk if they've not been used in n days
Archive CbrainTasks when they are old (need archiving system for tasks first, see issue #1212)
Lock user accounts that are no longer used
Check or probe DataProvider SSH tunnels (needed: API for that)
Notify sysadmins in real time about situations using email, twitter, etc.
- Disk filled
- Bourreau down
- Internal error messages
- User feedback received

Related: 1212, 1751, 1627 || #38 , 2086 || #39 , 2174 || #41 , 3128, 4987 || #42

Apr 14 '15 15:04 natacha-beck

Other feature: when batch-processing records of a particular type, highlight the rows of the affected record in the index page. Suggested in #297

Jan 15 '16 20:01 prioux

Maybe we can use the Raisl framework ActiveJobs :

http://guides.rubyonrails.org/active_job_basics.html

Sep 11 '16 16:09 prioux