Create a super metric when you want to check the health of your environment, but you cannot find a suitable metric to perform the analysis.

Procedure

  1. From the left menu, click Configure and then click Super Metrics.
  2. Click Add .
    The Create Super Metric wizard opens.
  3. (Optional) To edit a super metric, click the vertical ellipsis next to the super metric and select Edit. You can also edit the super metric using the EDIT option in super metrics page.
  4. Enter a meaningful name for the super metric such as Worst VM CPU Usage (%) in the Name text box.
    Note: It is important that you have an intuitive name as it appears in dashboards, alerts, and reports. For meaningful names, always use space between words so that it is easier to read. Use title case for consistency with the out of the box metrics and add the unit at the end.
  5. Provide a brief summary of the super metric in the Description text box.
    Note: Information regarding the super metric, like why it was created and by whom can provide clarity and help you track your super metrics with ease.
  6. From the Object Types drop-down list, select the object to associate with the super metric and click Next.
  7. Create the formula for the super metric.
    For example, to add a super metric that captures the average CPU usage across all virtual machines in a cluster, perform the following steps.
    1. Select the function or operator. This selection helps combine the metric expression with operators and/or functions. In the super metric editor, enter avg and select the avg function.
      You can manually enter functions, operators, objects, object types, metrics, metrics types, property, and properties types in the text box and use the suggestive text to complete your super metric formula.

      Alternatively, select the function or operator from the Functions drop-down menu.

    2. To create a metric expression, enter Virtual and select Virtual Machine from the object type list.
    3. Add the metric type, enter usage, and select the CPU|Usage (%) metric from the metric type list.
      Note: The expression ends with depth=1 by default. If the expression ends with depth=1, that means that the metric is assigned to an object that is one level above virtual machines in the relationship chain. However, since this super metric is for a cluster which is two levels above virtual machine in the relationship chain, change the depth to 2.

      The depth can also be negative, this happens when you need to aggregate the parents of a child object. For example, when aggregating all the VMs in a datastore, the metric expression ends with depth=-1, because VM is a parent object of datastore. But, if you want to aggregate all the VMs at a Datastore Cluster level, you need to implement 2 super metrics. You cannot directly aggregate from VM to Datastore Cluster, because both are parents of a datastore. For a super metric to be valid, depth cannot be 0 (-1+1=0). Hence, you need to create the first super metric (with depth=-1) for the aggregate at the datastore level, and then build the second super metric based on the first (with depth = 1).

      The metric expression is created.
    4. To calculate the average CPU usage of powered on virtual machines in a cluster, you can add the where clause. Enter where=””.
      Note: The where clause cannot point to another object, but can point to a different metric in the same object. For example, you cannot count the number of VMs in a cluster with the CPU contention metric > SLA of that cluster. The phrase "SLA of that cluster " belongs to the cluster object, and not to the VM object. The right operand must also be a number and cannot be another super metric or variable. The where clause cannot be combined using AND, OR, NOT, which means you cannot have where="VM CPU>4 and VM RAM>16" in your super metric formula.
    5. Position the pointer between the quotation marks, enter Virtual, and select the Virtual Machine object type and the System|Powered ON metric type.
    6. To add the numeric value for the metric, enter ==1.
    7. To view hints and suggestions, click ctrl+space and select the objects, object types, metrics, metrics types, property, and properties types to build your super metric formula.
    8. Select This option from the drop-down menu.

      If This option is selected during the creation of a metric expression, it means that the metric expression is associated to the object for which the super metric is created.

  8. Select the unit of the super metrics from the Unit drop-down.
    Note: The super metrics unit configured here can be changed in the metrics charts, widgets, and views.
  9. Click Validate to verify that the super metric formula has been created correctly.
    After you click the Preview button the system selects a random object and displays a metric graph showing values for the current super metric. For example, if you have selected Host System in the Object Types tab, after you click the Preview button it will randomly select a host system object from the list of the available objects and displays the graph for the selected host. Alternatively, you can also type in the object name in the Object text box, and the result will also depend on the pre-selected object type.
    1. Expand the Preview section.
      A metric graph is displayed showing values of the metric collected for the object. Verify that the graph shows values over time.
    2. Click Next.
    The Policies page is displayed.
  10. Select the policy which you want to associate with super metric and click Update.
    The selected policy is applied to the super metric. You can view the super metric you created, the associated object type, and policy on the Super Metrics page.