Jobs, bulk actions and rollback

A job is one piece of work aimed at one site, a client group, or every site in the workspace. Eight kinds:

Job	What it does
Update every plugin	everything with an update waiting, asked of the site as the step runs
Update one plugin	one slug, on every site named
Install a plugin	one slug, from the WordPress directory
Switch a plugin on	activates it for every visitor of every site named
Switch a plugin off	deactivates it, and whatever it was doing stops
Switch the theme	changes the whole face of every site named
Refresh the caches	flushes the object cache and rebuilds rewrite rules
Run a scheduled event	on the plans that carry automation

The job page refreshes itself every few seconds while the job is going, pauses while the tab is hidden, and gives up after thirty minutes. Each site has a row: Done, Failed or Waiting, a try count, a time, and what the site actually said. Cancel stops what has not started yet. Retry queues a new job for the failures alone.

Rollback

Every plugin update is two steps per site: take a snapshot of that plugin, then update it. No snapshot, no update. If the snapshot fails, that site is not updated and the row says so. A site whose ReDock Connect is too old to make one is skipped rather than updated.

The snapshot sits on the plugin’s row with the version it holds. Press it and the restore is queued through the same queue, watched the same way and recorded with your name on it.

This is rollback for a plugin update, not a backup service. It holds the plugin it replaced, not your database, so keep whatever backup your host or your own tooling gives you.