Example Performance Analysis Monitors

In this article

    Virtual Machine Memory Ballooning and Swapping Analysis

    Monitor/ Alert Name

    Target

    Parent Monitor

    Default Threshold(s)

    State changes

    Alert Description

    Virtual Machine Memory Ballooning and Swapping Analysis

    VMGUEST

    Performance

    memoryBalloonPct > 50

    memorySwapRate >100

    cpuSwapWaitPct>25

     

    If one metric is >threshold, then state=Warn.

    If more than one metric is >threshold, then state=Critical

    The virtual machine [vm name] running on host [hostname] has a Memory resource issue.

    The alert description will be automatically generated when the monitor triggers, and will factor in the three metric and their relative values, and also other related metrics and configuration settings for the target object.

    For this particular monitor, the alert description may include one or more of the following entries:

    • If memory ballooning is an issue:

    ‘The VM has ballooned [memoryBalloonPct]% of total memory. Although ballooning is a normal function of memory sharing, high ballooning can cause excessive pagefile swapping within the VM.’

    • If host use of the disk swapfile is an issue:

    ‘The vSphere Host has swapped [memorySwapRate] KB/sec to disk for this VM.’

    • If the VM is in a wait state while host swapping occurs:

    ‘The VM has spent [cpuSwapWaitPct]% of runtime in a wait state while the vSphere Host swaps memory.’

    • Additional guidance to the root cause can also be included:

    ‘…Disk swapfile use indicates the Host is under memory stress, and reclaiming memory through better methods (such as ballooning) is failing.’

    • Additional troubleshooting guidance includes VMware Tools status for this VM:

    ‘Check that VMware Tools is installed in this VM as it is required for efficient vSphere Host memory management.’

    Datastore Unknown Files Analysis

    Monitor/ Alert Name

    Target

    Parent Monitor

    Threshold(s)

    Alert Description

    Veeam VMware: Datastore Unknown Files Analysis

    SHARED DATASTORE

    Configuration

    unknownFilesGB>50, warning

     

     

    Datastore [datastore name] contains files unknown to vCenter totaling [unknownFilesGB] GB in size.

    These files could be orphaned snapshots, orphaned VMs, or manually copied files such as ISO.

    Run the Diagnostic Task against the Datastore to see the full list of unknown files.

    Datastore is currently [usedPct]% used with [freeGB] GB space remaining.

    The monitor works as follows:

    • Collector posts the unknownFilesGB metric for each datastore.

    unknownFilesGB = [total datastore space usage] – [space used by known VM datastore files], in GB

      • Known files = VMDKs, snapshots, swapfiles, all VM files registered in a vCenter Server.
      • Therefore, unknown files = total of orphaned snapshots, orphaned VMs, manually copied ISOs, and so on.
    • The unknownFilesGB metric is obtained by Collector querying a vCenter Server; space used by known VM datastore files is calculated using the ‘committed storage’ property for all VMs, and summing these properties for each datastore.
    • There are two advanced Collector Settings in Virtualization Extensions UI:
      • EnableUnknownFilesScan (default True). If False, the Collector will not collect the unknownFilesGB metric for the datastore;
      • DatastoreScanInterval (default 12). The Collector will run a query every N intervals. Default is 12 intervals x 15 = 180 minutes (every 3 hours).
    • When the Veeam VMware: Datastore Unknown Files Analysis monitor is fired by a threshold above, the Veeam VMware: Datastore Unknown Files Analysis Diagnostic Task becomes available in Health Explorer state change events.

    You can also run the corresponding Scan Datastore for Unknown Files task from the Tasks sidebar for the selected Datastore. To learn more about the task, see Scan Datastore for Unknown Files Task.