Shadow, fix test for 30 seconds

Backlog has been cleared, but I let this task open, as there is some actionable to follow-up, like write a postmortem.

Decrease a little bit the threshold, I've seen several times ffmpeg processes becoming quasi-zombies at 0:38 CPU time

Use MAX_TIME

Fix shadow use issues

To detect easily this situation in the future we can use this nagios check:
https://github.com/wanelo/nagios-checks/blob/master/check_sidekiq_queue

There are big ffmpeg video rendering tasks currently:

I restarted Sidekiq, it's currently processing new jobs. I'm watching the queues evolution, processed count increases, enqueued is currently stable.

One of the problem is the lack of integration with Phabricator.

EOL

Reducing a little bit the priority, all works fine with the current config, and each is correctly documented.

On another Salt configuration, Woods Cloud, where I'm going to reuse some blocks initially prepared from Nasqueron rOPS, I'm picking 9002as it's the same than the UID.