diff options
| author | Gabriel Filion <gabriel@koumbit.org> | 2011-04-27 15:46:58 -0400 |
|---|---|---|
| committer | Gabriel Filion <lelutin@gmail.com> | 2012-04-10 00:09:15 -0400 |
| commit | 7fb80fbecfa5a186846a2909bd1d6ada32ba21e1 (patch) | |
| tree | 9ed1d2bf91f5914d2fad1c321c3247c3061f7757 | |
| parent | dddff2316ee0ce0dab3540538de44a3829cc5c07 (diff) | |
Define check_load as a function of the nb of cpus
The "check_load" nagios check should not be hardcoded to some arbitrary
value. The thresholds should rather be adjusted in function of how many
CPUs are present in the server.
The factors were chosen by prior experience but could be adjusted if
needed.
Factors for warning thresholds (where N is the nb of cpus):
* 1min : 5 * N
* 5mins : 4 * N
* 15mins : 3 * N
Factors for critical thresholds:
* 1min : 10 * N
* 5mins : 9 * N
* 15mins : 8 * N
| -rw-r--r-- | files/nrpe/nrpe_commands.cfg | 1 | ||||
| -rw-r--r-- | manifests/nrpe/base.pp | 14 |
2 files changed, 13 insertions, 2 deletions
diff --git a/files/nrpe/nrpe_commands.cfg b/files/nrpe/nrpe_commands.cfg index 1a26cc3..b725a36 100644 --- a/files/nrpe/nrpe_commands.cfg +++ b/files/nrpe/nrpe_commands.cfg @@ -1,6 +1,5 @@ # default plugins command[check_users]=/usr/lib/nagios/plugins/check_users $ARG1$ -command[check_load]=/usr/lib/nagios/plugins/check_load $ARG1$ command[check_disk]=/usr/lib/nagios/plugins/check_disk $ARG1$ -x "/dev" -x "/dev/shm" -x "/lib/init/rw" command[check_apt]=sudo /usr/lib/nagios/plugins/check_apt -u $ARG1$ command[check_swap]=/usr/lib/nagios/plugins/check_swap $ARG1$ diff --git a/manifests/nrpe/base.pp b/manifests/nrpe/base.pp index 8636760..1d26849 100644 --- a/manifests/nrpe/base.pp +++ b/manifests/nrpe/base.pp @@ -23,9 +23,21 @@ class nagios::nrpe::base { # default commands nagios::nrpe::command { "basic_nrpe": - source => [ "puppet:///modules/site-nagios/configs/nrpe/nrpe_commands.cfg", + source => [ "puppet:///modules/site-nagios/configs/nrpe/nrpe_commands.${fqdn}.cfg", + "puppet:///modules/site-nagios/configs/nrpe/nrpe_commands.cfg", "puppet:///modules/nagios/nrpe/nrpe_commands.cfg" ], } + # the check for load should be customized for each server based on number + # of CPUs and the type of activity. + $warning_1_threshold = 5 * $processorcount + $warning_5_threshold = 4 * $processorcount + $warning_15_threshold = 3 * $processorcount + $critical_1_threshold = 10 * $processorcount + $critical_5_threshold = 9 * $processorcount + $critical_15_threshold = 8 * $processorcount + nagios::nrpe::command { "check_load": + command_line => "/usr/lib/nagios/plugins/check_load -w ${warning_1_threshold},${warning_5_threshold},${warning_15_threshold} -c ${critical_1_threshold},${critical_5_threshold},${critical_15_threshold}", + } service { "nagios-nrpe-server": ensure => running, |