Notification behavior out of the timeperiod

Icinga 2
1

Hi Folks,
I try to understand the notification behavior out of the timeperiod.

If the notification timeperiod is set from 12:00 p.m. to 12:03 p.m for a check and the users timeperiod is set to 24 hours with settings:

  states = [ OK, Critical ]
  types = [ Problem, Acknowledgement, Recovery, Custom, FlappingStart, FlappingEnd  ]

What should be the expected behavior as the check gets critical at 12:01 p.m. and recovers at 12:04 p.m.?

The user should get notified at 12:01 p.m. for the critical and skipped for the 12:04 p.m. recovery message. So far this is the expected behavior and its also working this way already.

But what should happen the next day at 12:00 p.m.? Do the user should get the skipped recovery message from the day before or not?

I ask, because I have two different Icinga2 instances which are both running on icinga2-2.12.3-1.el7.icinga.x86_64 and icingaweb2-2.8.2-1.el7.icinga.noarch and using the same configuration but I experience two different behaviors.

At one instance I get the recovery message the next day at 12:00 p.m. and on the other instance the recovery message is skipped fully. So I’m not sure if its a bug or where to search for the difference. First I would like to know, what is the expected behavior?

Many thanks in advance!

Best regards
David

2

Here is a example out of the check history:

history2
history2808×924 143 KB

Look at 26. April, check gets critical and sends out a critical message, while sending at the same time the OK from yesterday (I guess).

@dnsmichi

Is it a expected behavior or a bug? Sometimes the OK notification message is not send out and sometimes it is. So its not a constant behavior.

For example from 21. April till 22. April it worked as I would expect it:

history4
history4790×885 43.1 KB

Many thanks in advance!

Best Regards
David