Hello,
in our monitoring setup we are opening tickets in our issue tracker for problems using an API. When a problem resolves and sends a recovery notification those tickets are closed automatically, also using the API. This works perfectly fine and as intended for a very long time now.
Now the problem we are facing currently is, that some recoveries do not send a recovery notification, when the host/service started flapping after the initial alert for a problematic state was sent.
The following screenshot is one of many examples:
This leads to lots of overhead for us, since we do have quite a lot of tickets each day, where these issues need to be checked and resolved manually each time.
I have found the attribute “flapping_ignore_states” in the documentation, which says that it is configurable to for example ignore an OK state during flapping:
https://icinga.com/docs/icinga-2/latest/doc/08-advanced-topics/#check-flapping
Would this be a possible way to mitigate this problem? Or are there other alternatives also?
One more side note: If the attribute “flapping_ignore_states” would be a viable option, how could I integrate it into the Icinga Director, since the Director does not have an option to set this attribute?