To get a more accurate picture of the availability of your IT-services, it's a good idea to test the IT-services over the network-link that serves your centralized IT-services to other geographic locations.
This how-to paper will describe how to perform the most basic test, a test that checks the average roundtrip-time and measures packet loss using check_ping. The test will be performed by a remote Microsoft Windows server and called via check_nrpe and NSClient++. This test/check/service can then be added to a service group that you can use when generating SLA-reports.
Other, higher level test, can be performed in the same way. For example check_tcp or check_http can be used.
Configuration at your remote windows server
- Download and install .net framework 2.0 or later.
- Download latest binary version of nagiospluginsnt:
- Place the plugins in some directory, for example C:\plugins
- Testrun check_ping.exe at command-line. A good idea is to perform the test against a couple of the servers providing your centralized IT-services.
C:>C:\plugins\check_ping.exe -H mailserver.domain.com -w 100,20% -c 500,60%
- Add a command to NSClient++'s configuration file custom.ini:
command[check_ping]=C:\plugins\check_ping.exe -H $ARG1$ -w $ARG2$ -c $ARG3$
If you want the possibility to supply arguments, like in the above example, you must explicitly allow this using the setting allow_arguments=1 in op5.ini.
In previous versions of check-nt the argument-numbers are numbered with number one starting directly after the plugin-name (in the above example ARG1 is "-H" ARG2 is "mailserver.domain.com"), this is corrected in newer versions of check_nrpe. In newer versions ARG1 is always ARG1. If you're using old plugins, then this command is the one to put in op5.ini: command[check_ping]=C:\plugins\check_ping.exe -H $ARG6$ -w $ARG7$ -c $ARG8$
- Restart NSClient++ to load the new settings.
Configuration at your op5 Monitor server
- Test your new nrpe command at command-line:
monitor!monitor:~$ /opt/plugins/check_nrpe -H remote-windows-server.domain.com -c check_ping -a mailserver.domain.com 100,20% 500,60%
- Add a check-command to your configuration:
command_line: $USER1$/check_nrpe -H $ARG1$ -c check_ping -a $ARG2$ $ARG3$ $ARG4$
- Add a new service to the host that you want to test the responsetime to:
service_name: Responsetime from location