[naemon-users] Keep acknowledgements on unknown/timeout flaps

[Christian Kujau]

On Fri, 23 Jul 2021, Florian Panzer - PLUSTECH GmbH wrote:
> I don't really want to do both of them. If i choose option B, I won't be
> notified if the status changes to CRITICAL.

Did you play around with the notification_options[0] setting? If set to 
"c,w,r" notifications are not sent on UNKNOWN states at all:

define service {
        service_description     check_test
        host_name               foobar
        use                     remote-service
        check_command           check_dummy!1
        notification_options    c,w,r
        max_check_attempts      1
}

Flap detection[1] options could also help here, if it's only going into 
UNKNOWN for a short time.

The more I think about it, the more options come to mind: retry_interval 
could be used to increase the time until the check goes to UNKNOWN at all 
(IIRC), thus preventing an unwanted UNKNOWN notification. And 
notification_period could be used to (not) send out notifications during a 
specifc timeframe.

HTH,
C.

[0] https://www.naemon.org/documentation/usersguide/objectdefinitions.html#service
[1] https://www.naemon.org/documentation/usersguide/flapping.html
-- 
BOFH excuse #62:

need to wrap system in aluminum foil to fix problem