Testing the notification system

[nicole]

Hi

I know that this might not be 100% Zabbix related, but since many of you probably did something similar, I would like to ask you some ideas for testing the Zabbix notification system.

Given that Zabbix can be installed and used in several different kinds of environment, some personalized testing might be required by each administrator. In my case, there's nothing too fancy to be done. All I need to do is to monitor CPU usage, disk space availability and memory usage (with notifications via email if those metrics are above or below a given percentage I haven't defined yet).

Problem is I cannot just create a bunch of processes and files on my server to verify if Zabbix will actually notify me about CPU, memory and disk. I tested the email notifications on a local machine for other metrics, and I can receive emails through Zabbix, so I can safely assume the configuration is correct.

So, how you Zabbix people test your notification system? Do you just trust that it will work when something go wrong or do you create a proper test environment (and, if so, how exactly you did that)?

[guzzijason]

One thing you can do is temporarily change the thresholds on your triggers to something you know will activate it. For example, if you have a disk usage trigger that is set to go off at 70%, change it to something really low like 10% to force the trigger to go off. Once you're happy that it worked as intended, then change it back to the real threshold of 70%.

__Jason