In this scenario, you investigate and resolve the Virtual machine has an unexpected high CPU workload alert. The alert might be generated for more than one virtual machine.

Prerequisites

Generated alerts in vRealize Operations appear in the alert lists. You use the alert lists to investigate, resolve, and begin troubleshooting problems in your environment.

Procedure

  1. From the left menu, click Troubleshoot and then click Alerts.
  2. To limit the list to virtual machine alerts, click All Filters on the toolbar.
    1. Select Object Type in the drop-down menu.
    2. Enter virtual machine in the text box.
    3. Click Enter.
    The alerts list displays only alerts based on virtual machines.
  3. To locate the alerts by name, enter high CPU workload in the Quick filter (Alert) text box.
  4. In the list, click the Virtual machine has an unexpected high CPU workload alert name.
  5. Review the information. To show the recommendations, click Configuration > Recommendations in the left pane .
    Option Evaluation Process
    Alert Description Review the description so that you better understand the alert.
    Recommendations Do you think that implementing one or more of the recommendations can resolve the alert?
    What is Causing the Issue?

    Do the triggered symptoms support the recommendations? Do the other triggered symptoms contradict the recommendation, indicating that you must investigate further?

    In this example, the triggered symptoms indicate that the virtual machine CPU demand is at a critical level and that the virtual machine anomaly is starting to get high.

    Non-Triggered Symptoms

    Some alerts are generated only when all the symptoms are triggered. Others are configured to generate an alert when any one of the symptoms are triggered. If you have non-triggered symptoms, evaluate them in the context of the triggered alerts.

    Do the non-triggered symptoms support the recommendations? Do the non-triggered symptoms indicate that recommendations are not valid and that you must investigate further?

  6. To resolve the alert based on the recommendation to check the guest applications to determine whether a high CPU workload is an expected behavior, click the Action menu on the center pane toolbar and select Open Virtual Machine in vSphere Client.
    1. Log in to the vCenter Server instance using your vSphere credentials.
    2. Start the console for the virtual machine and identify which guest applications are consuming CPU resources.
  7. To resolve the alert based on the recommendation to add more CPU capacity to this virtual machine, click Set CPU Count for VM.
    1. Enter a new value in the New CPU text box.

      The value that appears is the calculated suggested size. If vRealize Operations was monitoring the virtual machine for six or more hours, depending on your environment, the value that appears is the CPU recommended size metric.

    2. To allow power off or to create a snapshot, depending on how your virtual machines are configured, select the following options.
      Option Description
      Power Off Allowed

      Shuts down or powers off the virtual machine before modifying the value. If VMware Tools is installed and running, the virtual machine is shut down. If VMware Tools is not installed or not running, the virtual machine is powered off without any regard for the state of the operating system.

      In addition to the question whether the action shuts down or powers off a virtual machine, you must also consider whether the object is powered on and what settings are applied.

      Snapshot

      Creates a snapshot of the virtual machine before you add CPUs.

      If the CPU is changed with CPU Hot Plug enabled, then the snapshot is taken with the virtual machine running, which consumes more disk space.

    3. Click OK.
    The action adds the suggested number of CPUs to the target virtual machine.
  8. Allow several collection cycles to run after implementing the suggested changes and check the alert list.

What to do next

If the alert does not reappear after several collection cycles, it is resolved. If it reappears, further troubleshooting is required.