Fix that no_more_notifications gets reset when Recovery notifications are filtered away

[dnsmichi]

Hello from the OSMC hackathon :)

Description

Combination of notification interval = 0, times.begin = 1h and filtered recovery notificadtions.

The problem arises with the first critical/down notification which delays the notification to 3h hours in the future, and sets no_more_notifications to true.

Once a recovery notification is triggered, this does not reach the point where it actually evaluates users to notify. It also does not reset no_more_notifications to false again, allowing our NOTOK-OK-NOTOK state transition logic to kick in.

We've analysed the problem with @jschanz and created a possible fix for it. A workaround for the problem is to add the Recovery type again to the notification objects and send a dummy OK check results to also reset the no_more_notifications flag to false again.

[jschanz]

[dnsmichi]

Since there's a workaround for now, I'd like to see the fix being tested a bit more before backporting it to 2.10. Therefore scheduling for 2.11 or 2.10.x later.