Skip to main content
Active IQ Unified Manager 9.7
9.7
A newer release of this product is available.

Monitoring virtual machines

Contributors

For any latency issue on the virtual machine (VM) applications, you might need to monitor the VMs to analyze and troubleshoot the cause.

Active IQ Unified Manager enables you to view the detailed topology of a VM application that displays the components to which the VM is related, for example, the host of the VM or the datastore connected to it. The topology view displays the underlying components in their specific layers, in the following order: Virtual Disk > VM > Host > Network > Datastore > VMDK.

Note

For viewing the metrics of the datastore (in the storage layer), the parent ONTAP cluster must be added to Unified Manager, and the discovery (polling or metrics collection) for the cluster must be complete.

You can determine the I/O path and component-level latencies from a topological aspect and identify whether storage is the cause of the performance issue. The summary view of the topology displays the I/O path, and highlights the component that has IOPS and latency issues for you to decide on the troubleshooting steps. You can also have an expanded view of the topology that depicts each component separately along with latency of that component. You can select a component to determine the I/O path highlighted through the layers.

vCenter data collection schedule

For polling vCenter configuration data, Unified Manager follows the same schedule as for collecting cluster configuration data. For information about vCenter configuration and performance data collection schedule, see "Cluster configuration and performance data collection activity".

vCenter collects real-time 20-second performance data samples and rolls them up to 5-minute samples. The schedule for performance data collection of Unified Manager is based on the default settings of vCenter server. Unified Manager processes the 5-minute samples obtained from vCenter and calculates an hourly average of the IOPS and latency for the virtual disks, VMs, and hosts. For datastores, Unified Manager calculates an hourly average of the IOPS and latency from samples obtained from ONTAP. These values are available at the top of the hour. The performance metrics are not available immediately after a vCenter is added, and is available only when the next hour starts. Performance data polling begins on completing a cycle of configuration data collection.

Viewing summary topology

To determine performance issues by viewing the VMs in a summary topology

  1. Go to VMWARE > Virtual Machines.

  2. Search for your VM by typing its name in the search box. You can also filter your search results based on specific criteria by clicking on the Filter button. However, If you cannot find your VM, ensure that the corresponding vCenter has been added and discovered.

    Note

    vCenter servers allow special characters (such as %, &, *, $, #, @, !, \, /, :, *, ?, "`, <, >, |, ;, ') in the names of vSphere entities, such as VM, cluster, datastore, folder, or file. The VMware vCenter Server and ESX/ESXi Server do not escape special characters used in the display names. However, when the name is processed in Unified Manager, it is displayed differently. For example, a VM named as %$VC_AIQUM_clone_191124% in vCenter server is displayed as %25$VC_AIQUM_clone_191124%25 in Unified Manager. You must keep a note of this issue when you query a VM with a name having a special characters in it.

  3. Check the status of the VM. The VM statuses are retrieved from vCenter. The following statuses are available. For more information about these statuses, see VMware documentation.

    • Normal

    • Warning

    • Alert

    • Not monitored

    • Unknown

  4. Click the down arrow beside the VM to see the summary view of the topology of the components across the compute, network, and storage layers. The node that has latency issues is highlighted. The summary view displays the worst latency of the components. For example, if a VM has more than one virtual disk, this view displays the virtual disk that has the worst latency among all the virtual disks.

  5. To analyze the latency and throughput of the datastore over a period of time, click the Workload Analyzer button on top of the datastore object icon. You go to the Workload Analysis page, where you can select a time range and view the performance charts of the datastore. For more information about workload analyzer, see Troubleshooting workloads using the workload analyzer.

Viewing expanded topology

You can drill down to each component separately by viewing the expanded topology of the VM.

  1. From the summary topology view, click Expand Topology. You can see the detailed topology of each component separately with the latency numbers for each object. If there are multiple nodes in a category, for example multiple nodes in the datastore or VMDK, the node with worst latency is highlighted in red.

  2. To check the IO path of a specific object, click on that object to see the IO path and the corresponding mapping. For example, to see the mapping of a virtual disk, click the virtual disk to view its highlighted mapping to the respective VMDK. In case of a performance lag of these components, you can collect more data from ONTAP and troubleshoot the issue.

    Note

    Metrics are not reported for VMDKs. In the topology, only the VMDK names are displayed, and not metrics.