You use different policy requirements for your Development, Test, and Production environments so that you can configure the specific policy settings for vRealize Operations Manager to analyze and report on your objects, including your virtual SQL Servers.

About this task

This scenario presents several typical cases where you might be required to differentiate between the policy requirements for Development, Test, and Production environments.

  • For your Development and Test environments, you might not be concerned if the objects in these environments experience network redundancy loss, but you do care when the objects fail. In this case, you locate the Physical NIC link state alert definition, double-click the state, and set it to Disabled.

  • For a Test environment, you might not be concerned if your virtual machines demand more memory and CPU capacity than what is actually configured, because workloads can vary in test environments.

  • For a Production environment, your virtual machines might require more memory than you have configured, which might cause a problem with the performance and reliability of your production environment.

In this procedure, you override the symptom definition threshold value for the Co-Stop performance of your virtual machines.

Prerequisites

Verify that the following conditions are met:

  • You created a custom policy for your virtual SQL Servers. See Create a Policy to Meet vSphere Operational Needs.

  • You understand the Co-Stop CPU performance metric for virtual machines. This metric represents the percentage of time that a virtual machine is ready to run, but experiences delay because of co-virtual CPU scheduling contention. Co-Stop is one of several performance metrics for virtual machines that also include Run, Wait, and Ready.

  • The alert definition named Virtual machine has high CPU contention caused by Co-Stop, exists.

  • Symptom definitions exist to track the critical, immediate, and warning levels of CPU Co-Stop on the virtual machines. For example, the critical level for virtual machine CPUs that experience contention more than 15% of the time is set to 15% by default, as measured by the Co-Stop metric. The default threshold level for Immediate is 10%, and for warning is 5%. However, in your production policy for your production virtual machines, you manage the critical level at 3%.

Procedure

  1. On the Policy Library tab, locate your vSphere Production Virtual SQL Servers policy, and click the pencil to edit the policy.

    The Edit Monitoring Policy workspace appears.

  2. In the workspace, click Override Alert / Symptom Definitions.
  3. On the Alert Definitions pane, enable the Co-Stop alert definition to notify you about high CPU contention on your virtual machines.
    1. In the Object Type drop-down menu, select vCenter Adapter and Virtual Machine.
    2. In the Search text box, enter stop to display only the alert definitions that relate to the Co-Stop performance metric for virtual machines.
    3. For the Alert definition named Virtual machine has high CPU contention caused by Co-Stop, click the State drop-down menu and click Enabled.
  4. In the Symptom Definitions pane, modify the critical Co-Stop level for virtual machines so that vRealize Operations Manager triggers an alert based on the threshold level defined for this symptom.
    1. In the Object Type drop-down menu, click vCenter Adapter and Virtual Machine.
    2. In the Search text box, enter stop to display the symptom definitions that apply to the Co-Stop performance metric for virtual machines.
    3. For the symptom definition named Virtual Machine CPU Co-stop is at Critical level, click the State drop-down menu and click Enabled.
    4. Click the Condition drop-down menu, and click Override.

      For a production policy, a typical critical threshold value is >3. For a development or test environment policy, a typical critical threshold value is >10.

    5. In the Override Symptom Definition Threshold dialog box, enter >3 to change the threshold value, and click Apply.
  5. Modify the immediate Co-Stop level for virtual machines.
    1. For the symptom definition named Virtual Machine CPU Co-stop is at Immediate level, click the State drop-down menu and click Enabled.
    2. Click the Condition drop-down menu, and click Override.
    3. In the Override Symptom Definition Threshold dialog box, enter >2 to change the threshold value, and click Apply.
  6. Modify the warning Co-Stop level for virtual machines.
    1. For the symptom definition named Virtual Machine CPU Co-stop is at Warning level, click the State drop-down menu and click Enabled.
    2. Click the Condition drop-down menu, and click Override.
    3. In the Override Symptom Definition Threshold dialog box, enter >1 to change the threshold value, and click Apply.
  7. Click Save to save your policy.

Results

You changed the Co-Stop CPU performance metric for virtual machines to minimize the delay on your SQL Server virtual machines because of CPU scheduling contention.

What to do next

Create a group type to use to categorize your group of virtual SQL Servers, create an object group that contains your virtual SQL Servers, and apply the policy to your object group.