Out of the box vROps comes with hundreds of Alerts for VMs, one of them is focused on disk space, namely Windows Drives and Linux Filesystems.
Selecting the Alert shows the details, including the Symptoms.
The Symptoms are the actual triggers of the Alert, in this case there are two Symptoms: Guest File System Partition beyond 95% (Critical) and another beyond 90% (Warning). By Guest File System Partition we mean Windows Drive or Linux Filesystem. Let's look at the Symptoms.
These are the two Symptoms being used in the original Alert. Let's look at the details of the Critical one.
The Symptom definition gives us the details: name, object, criticality, wait/cancel cycle, threshold type, what Alert its being used in, and what Policies its enabled in. Selecting the ellipses allows you to edit the Symptom.
Here we get more details and the ability to evaluate on instanced metrics. This means that the Symptom will be evaluated on each instance of the Guest File System Partition, which is to say each Windows Drive and each Linux Filesystem. What if you don't want to monitor every Windows Drive and every Linux Fileystem? Well, we can exclude certain Drives/Filesystems via the box at the bottom. Here's how you do it.
First clone the Symptoms and Alerts and make adjustments to exclude the Windows Drive we want. Go back to Configure - Alerts - Symptom Definitions nd clone the original Symptoms, let's start with the Critical one.
Let's exclude the H:\ Drive from this Symptom, but you could do the same for any Windows drive. Not every metric on every VM will show by default in the list of metrics, so you may have to search for a VM.
Here I've chosen a Windows VM with an H:\ drive, dragged it over to the exclusion box, given my Symptom a descriptive name, and clicked SAVE. Next, let's clone our Alert to use that new Symptom.
Give it a new name, something descriptive. In our case we're creating Alerts/Symptoms for Windows Servers excluding the H:\ drive, we don't want to monitor that drive as it's always full in our case.
Click NEXT.
Let's remove the inherited Symptoms and add our newly created one excluding the H:\ Drive.
Click NEXT, give it a Recommendation if you want, click NEXT, don't enable it in any Policies, click UPDATE.
Next let's create a new Policy, that is bound to just our Windows VMs, that contains this Alert. First create a Custom Group for Windows VMs. Go to Environment - Custom Groups - ADD, and create it like this.
You can always click PREVIEW to confirm it contains just the Windows VMs, or whatever you're trying to create a group for. Next let's create a Policy for that Custom Group with our new Alert in it. Go to Configure - Policies - ADD and create your new Policy.
Next let's enable the Alert, Bind the Windows VMs Group, and enable the policy. Click Alerts and Symptoms - Alert Definitions - Virtual Machine, search for your new Alert, and Enable it.
Be sure to disable the original alert in your new Policy, in this case the "One or more virtual machine guest file systems are running out of disk space". This ensures that the new Alert effectively replaces the old one for the Custom Group of VMs bound to this policy.
Next click Groups and Objects and bind the Policy to the Windows VMs Group.
Back at the Policies page you'll notice it has 1 Locally Defined Alert and 1 Associated Custom Group.
The Policies dropdown now shows our new Policy, which is a child Policy of the vSphere Default Policy. This new Policy has Priority 1, which means objects will run through it first, then through the vSphere Default Policy which has Priority D (Default), which is the catch all Policy. If nothing else, each object will hit this Policy.
For more on vROps Policies go here. We will discuss them in a future blog. At this point, all drives besides the H:\ drive are being monitored on all Windows VMs. This same methodology will work for Linux Filesystems. For more information and a trial of vROps go here!