Nagios: Possible to be alerted between certain times, and not when notification_period resumes

icinganagios

I want to check a service 24×7, but want to check certain thresholds only during business hours. I could use check_period, but that would mean that the services aren't checked outside business hours. Alternatively I could use notification_period, but that will mean that when the notification_period starts, any alerts will be sent then, and I don't want that.

Is there any way of achieving this with Nagios 3 (actually I'm using icinga)?

Best Answer

I can think of two ways to do this: (a) use an external command to change the check command (Nagios calls this "adaptive monitoring") or (b) split the service into two with different check commands and periods.

I'll use check_load as and example with these (skeletal) service and command definitions:

 define service{
   name          load
   host_name     foohost
   check_command check_load!1,1,1!2,2,2
   ... (all other options)
 }

 define command{
   name         check_load
   command_line $USER1$/check_load -w $ARG1$ -c $ARG2$
 }

For (a) suppose you wish to change these values at 8pm return them at 8am. In cron add

 0 20 * * * /some/path/change_load_check 3,3,3 4,4,4
 0  8 * * * /some/path/change_load_check 1,1,1 2,2,2

where change_load_check is

#!/bin/sh

now=`date +%s`
commandfile='/usr/local/nagios/var/rw/nagios.cmd'

W=$1
C=$2

/bin/printf "[%lu] CHANGE_SVC_CHECK_COMMAND;foohost;load;check_load!$W!$C\n" \
  $now > $commandfile

You need to have external commands enabled.

For (b) you would take the original service, turn it into a template, and create two new services that specify different periods and check commands like so: 

 define service{
   name          load_template
   host_name     foohost
   ... (all other options)
   register      0
 }

 define service{
   name                load_workhours
   use                 load_template
   check_period        workhours
   notification_period workhours
   check_command       check_load!1,1,1!2,2,2
 }

 define service{
   name                load_offhours
   use                 load_template
   check_period        offhours
   notification_period offhours
   check_command       check_load!3,3,3!4,4,4
 }
Related Topic