Service Notification in SOFT State

Troubleshooting
1

Hello everybody,

we always have that individual services send a notification in the SOFT state.
The configurations “check attempts” or “notification delay” do not work.

However, I was able to determine that it occurs where the counter for “service alarm number” is not reset even though the status is OK again.
In my understanding, this should happen when the status changes from an error state to OK.

The only thing that works is the periodic alarm. Only after 30 minutes have passed will the alarm be given again.

OS and CMK version
CentOS 6.10
CMK 1.6.0p3

LQL output:
Host HOSTNOK has service_current_notification_number> 0 and reports immediately when the status changes. The host HOSTOK only when the type is HARD.
OMD [cmk_int]: ~ $ lq “GET services \ nColumns: host_name state service_current_notification_number \ nFilter: description = Check_MK \ nFilter: host_name = HOSTOK”
HOSTOK; 0; 0
OMD [cmk_int]: ~ $ lq “GET services \ nColumns: host_name state service_current_notification_number \ nFilter: description = Check_MK \ nFilter: host_name = HOSTNOK”
HOSTNOK; 0; 19

Thanks
Marc Wolff

2

First question enterprise core or Nagios?
To your livestatus queries. Why the “service_current_notification_number” if you search for hosts?
host_current_notification_number would be a better option here.
If this happens in your system then your core has a big problem. I have no system with this symptom.

3

Hi Andreas,
thank you for the replay.

We use the enterprise core.
We have seen the problem by services currently. And only by the services where the service_current_notification_number > 0

is some configuration possible that block the reset of this counter?

Thanks
Marc

4

Hello everybody,

i found the reason for my problem.
If “notified events for services” has been configured without “service recovers to ok”, the counter for service_current_notification_number is not reset

Thanks
Marc