Perl Monitor for Linux Volumes

Version 3

    We're not quite at the release where I can start using agents, and I suppose some folks might want another way of getting at this data without installing the new Linux agents, so I wanted to share this since it has solved a big problem for me.

     

    Anyone who has linux servers in their environment is probably familiar with how bad net-SNMP is when reporting volume utilization to Orion. Originally, we were just lowering our thresholds for linux to try and accommodate the delta in reporting, however this ultimately mis-represents the criticality of what's going on to NOC. I wrote this in response to a server getting to 99% utilization during non-business hours, giving the responding party just minutes to fix the storage issue before bigger problems started piling up.

     

    The perl script below will allow for two thresholds, one for normal volumes, and another for 'hot' volumes that typically run at higher utilization. There is also a variable to skip volumes if you want to ignore things that just don't matter.

     

    As always, I recommend you read through the script, and make some observations after deployment to confirm you're getting what you think you're getting. The regex I'm using on line 13 may need to be adjusted for your environment.

     

    The monitor will return the count of volumes exceeding your thresholds as your statistic, and your message will contain all of these volumes outside of threshold and their utilization. While this isn't very good for historical trending, it's great for capturing accurate volume alerts on Linux servers.