I was recently given a use case by a client who was seeing alert noise from the vR Ops 6.1 alert “One or more virtual machine guest file systems are running out of disk space“. This alert is built from three symptoms:
“1. Guest file system space usage at warning level”
“2. Guest file system space usage at immediate level”
“3. Guest file system space usage at critical level”
The logic in the alert is that “ANY” of these symptoms can be true to fire off the Alert.
For many organizations this is a reasonable out of the box default Alert setting. However For this client, and others I’m sure, it did not work and needed to be customized. For this team it was triggering for almost all of the windows machines in their environment.
When this team builds their windows machines, they put the page file on a separate disk The P: drive, sized to accommodate just the pagefile.sys file and not much else. So this disk on most of their windows machines is always at > 95% Usage. So of course the Out of the box alert storms the vR Ops Alert inbox.
This is how we addressed this challenge:
- Adjust the default policy to turn off the Out Of the box Alert “One or more virtual machine guest file systems are running out of disk space”. This will also turn off this alert for non windows machines. This team elected to follow up with another custom configuration for Linux machines.
- Build a Custom Group that uses a vSphere Tag to gather all the required Windows VMs.
- Build a set of Symptoms for each drive they want to alert on.
- Create a new alert or copy the existing Alert and base it on new Symptoms.
- Create New Policy and apply it to the custom group.
Enjoy Less Alerts, enjoy customized Alerts, and be more productive!!