vRealize Operations Manager Super Metrics

Super metrics are the equivalent of a custom metric, it can be used to help simplify tracking by watching a few key metrics instead of many metrics. By bringing forward related data, it helps to troubleshoot specific information, which speeds time to resolution. A super metric is a mathematical formula that contains one or more metrics.

After you create a super metrics, you assign it to one or more object types. For instance, you could assign a super metrics to cluster object (resource kinds) that calculates the average memory demand on all virtual machines in the cluster.

Super metrics can be found on the Content page. vRealize Operations Manager keeps a consistent format, click the + sign to add a new super metrics, the pencil to edit a super metrics, the two triangles to clone a super metrics, the x to delete a super metrics, and clicking the gear icon allows you to import and export super metrics.

When creating a super metric, there are several different functions to choose from for your formula. They are either looping functions or single functions. Looping functions work on more than one value and single functions on a single value.

Also, there are three different categories of super metrics. They included rollup, generic resource, and specific resource or pass-through.

• Rollup: The ability to apply a looping function (sum, avg, etc) to a group of resources that all have the same metrics.
• Generic Resource: The ability to combine metrics on a type of resource using standard mathematical operators.
• Specific resource or pass-through: The ability to combine specific resource metris and apply them to any resource.

The looping functions are:

• avg: Average of collected values
• combine: Combines all of the values of the metrics of the included objects in a single metric timeline
• count: Number of values collected
• max: Maximum value of the collected values
• min: Minimum value of the collected values
• sum: Total of the collected values

These are the core functions we are going to work with in this post. Other functions, such as atan (Arctangent of x) are inverse trigonometric functions that I have no intention of covering!

The last function that needs to be discussed is depth, depth supports multilayer rollups. The default depth is 1, vRealize Operations Manager searches for objects one level below the object type where you assign the specific metrics. If you have a positive number you go down the relationship chain and if you have a negative number you move up the relationship chain.

Let's build out a super metric to show the average CPU utilization for all the virtual machines on a host. First, we are going to select the looping function of avg from the Functions drop-down selector.

Next we are going to select the left parenthesis ( from the Operators drop-down, we are going to put our formula between the open and closed operators.

On the Object Types, we are going to drop down the Adapter Type drop-down list and select vCenter Adapter.

From the object types that appear, we are going to scroll down until we find Virtual Machine and select it. In the lower left panel, on the Metrics list, we are going to expand the CPU category, then scroll down and double-click the Usage (%) metric.

In the above picture, we can see that the formula shows in the top panel as avg(${adaptertype=VMWARE, objecttype=VirtualMachine, metric=cpu|usage_average, depth=1}). You need to remember after double-clicking Usage% metric, that we need to add the right parenthesis ) from the Operators drop-down to close the formula.

Give your super metric a name, you can see in my example I have used the name Average Virtual Machine CPU Usage %.

Next we are going to verify the super metric formula is working. Click the Visualize Super Metric button. On Object Types, click the Adapter Type drop-down menu and select vCenter Adapter. Next scroll down and select Host System, then select one of the child object hosts listed in the objects panel. If information is displayed in the visualization window, you know the formula is correct.

After you verified that your super metric works, we need to add it to an object. On the Objects Types tab we are going to click the + button to add a new object type.

Expand the vCenter Adapter object and then add Host Systems.

Before we can see our super metric in All Metrics on the Troubleshooting tab, we need to add it to the default policy. We are going to edit the vSphere Solution's Default Policy and go to Collect Metrics and Properties. 

Drop-down the Attribute Type filter and uncheck Metric and Property, this will bring up only the super metric you created.

Change the State of the super metrics from Inherited to Local and save your changes.

Our super metrics should now be available in All Metrics under Troubleshooting, because we set the Object Type as Host System, we will only see this super metrics when a host has been selected. In the picture below, I have expand the Super Metrics tree and Average Virtual Machine CPU Usage % is available for labesx07.home.virtlab.com.

After double-clicking the metrics, a chart for the average virtual machine CPU usage % on labesx07 displays on the right; below that chart is CPU usage of the two virtual machines on the host. The SCOM virtual machine is running between 4.14% and 4.76% and the VMware vCenter Server Production virtual machine is running between 8.43% and 9.58%. The virtual machine average for both is 6.45% at that specific point in time.

This is merely a sample of what you can do with super metrics. Super metrics are really powerful, they are certainly worth exploring to find out how it can enhance your monitoring capabilities.