If you have a lot of VMs running on your ESX hosts, especially VMs with a high vCPU count, you will likely run into the high CPU %READY problem. VMware vCenter counts the %READY metrics only for VMs and vCenter Operations then can only process these metrics. Out of the box it is not possible to get these metrics for a cluster or even a vCenter. But with vCOps Enterprise it is possible to define your own dashboards and Super Metrics that calculate new metrics based on already existing data. In this article I show you how to do it.
Table of Contents
The Super Metric
First you must log on to your vCenter Operations Website (the Custom UI). I don’t know if it is also possible with the vCenter plugin, but if you have a larger environment you will likely use the custom installation of vCenter Operations.
Next we open the Super Metric Editor. You can find the Link in the main menu under Environment > Advanced > Super Metrics > Super Metric Editor.
At the top of the window we can define a formula manually, but it’s easier to use the resource kind search on the right side. The left side shows the actual resources (VM’s , Hosts, vCenter etc.), but in most cases it makes no sense to work directly with a resource. As the name states resource kinds make the formula work for every resource of that kind.
Add the avgN function to the formula and choose “4″ as the second option for it. avgN will calculate the average of all CPU Ready (%) metrics four levels below the resource kind it is attached to. Because we later attach that Super Metric to all vCenter Server resources it will seek four levels below ( Datacenter > Cluster > ESX Host > VM ) to find the Ready (%) Metric. If you want to know that for a Cluster choose “2″ and attach the Super Metric to a Cluster.
In the Part Two of that article i show how to add the Super Metric to a package and attach it to the resources.