Home
  •  

Utilisation du processeur d'ordinateur hôte de System Center Virtual Machine Manager 2019+

Microsoft.SystemCenter.VirtualMachineManager.HostCPUUtilizationMonitor (UnitMonitor)

Analyse l'utilisation du processeur de l'ordinateur hôte

Knowledge Base article:

Résumé

Ce moniteur fait le suivi des performances du processeur de l'ordinateur hôte. L'utilisation du processeur a dépassé le seuil.

Causes

Les ordinateurs virtuels qui s'exécutent sur l'ordinateur hôte consomment une trop grande quantité de ressources processeur de l'ordinateur hôte.

Résolutions

Pour libérer des ressources processeur, procédez à la migration d'un ordinateur virtuel vers un autre ordinateur hôte. Vous pouvez également régler Virtual Machine Manager pour qu'il optimise automatiquement des clusters hôtes en activant la fonction Optimisation dynamique. Pour plus d'informations sur l'optimisation dynamique, consultez la documentation de Virtual Machine Manager.

Element properties:

TargetMicrosoft.SystemCenter.VirtualMachineManager.Discovery.HyperVHost
Parent MonitorSystem.Health.AvailabilityState
CategoryAvailabilityHealth
EnabledTrue
Alert GenerateTrue
Alert SeverityMatchMonitorHealth
Alert PriorityNormal
Alert Auto ResolveTrue
Monitor TypeMicrosoft.SystemCenter.VirtualMachineManager.Performance.PowershellBasedConsecutiveSamplesTwoThresholdsMonitorType
RemotableTrue
AccessibilityPublic
Alert Message
L'utilisation du processeur de l'ordinateur hôte est élevée.
L'utilisation du processeur d'ordinateur hôte est de {1} et a dépassé les niveaux d'avertissement ou critiques pour {0}
RunAsDefault

Source Code:

<UnitMonitor ID="Microsoft.SystemCenter.VirtualMachineManager.HostCPUUtilizationMonitor" Accessibility="Public" Target="VMMDiscovery!Microsoft.SystemCenter.VirtualMachineManager.Discovery.HyperVHost" ParentMonitorID="SystemHealth!System.Health.AvailabilityState" Remotable="true" TypeID="Microsoft.SystemCenter.VirtualMachineManager.Performance.PowershellBasedConsecutiveSamplesTwoThresholdsMonitorType">

  <Category>AvailabilityHealth</Category>

  <AlertSettings AlertMessage="Microsoft.SystemCenter.VirtualMachineManager.HostCPUUtilizationMonitor.AlertMessage">

    <AlertOnState>Warning</AlertOnState>

    <AutoResolve>true</AutoResolve>

    <AlertPriority>Normal</AlertPriority>

    <AlertSeverity>MatchMonitorHealth</AlertSeverity>

    <AlertParameters>

      <AlertParameter1>$Target/Host/Property[Type="Windows!Microsoft.Windows.Computer"]/NetworkName$</AlertParameter1>

      <AlertParameter2>$Data/Context/SampleValue$</AlertParameter2>

    </AlertParameters>

  </AlertSettings>

  <OperationalStates>

    <OperationalState ID="HostCPUUtilizationReachedWarning" MonitorTypeStateID="ConditionWarn" HealthState="Warning"/>

    <OperationalState ID="HostCPUUtilizationReachedError" MonitorTypeStateID="ConditionError" HealthState="Error"/>

    <OperationalState ID="HostCPUUtilizationIsSuccess" MonitorTypeStateID="ConditionSuccess" HealthState="Success"/>

  </OperationalStates>

  <Configuration>

    <IntervalSeconds>180</IntervalSeconds>

    <TimeoutSeconds>180</TimeoutSeconds>

    <WarningThreshold>90</WarningThreshold>

    <CriticalThreshold>95</CriticalThreshold>

    <NumSamples>3</NumSamples>

    <hostComputerName>$Target/Host/Property[Type="Windows!Microsoft.Windows.Computer"]/NetworkName$</hostComputerName>

    <ScriptBody><Script>#This Script is used to check the Host CPU utilization

# which includes parent partition and CPU utilizations caused by the vms 

# Alert when the utilization of this partition is high:

# Warning: WarningThreshold%, like 80%

# Critical: CriticalThreshold, like 90%

#SetLocale("en-us")

#Please check this link to get understanding http://blogs.msdn.com/b/tvoellm/archive/2008/07/14/how-to-get-processor-utilization-for-hyper-v-via-wmi.aspx



$momApi = new-Object -comObject "Mom.ScriptAPI"

$momBag = $momApi.CreatePropertyBag()

$timePeriodSeconds = 10 

#$strLPTotalQuery = "SELECT * FROM Win32_PerfRawData_HvStats_HyperVHypervisorLogicalProcessor where Name='_Total'"

#$strLPCountQuery = "SELECT * FROM Win32_PerfRawData_HvStats_HyperVHypervisorLogicalProcessor"

$logicalProcessorStats = Get-CimInstance -ClassName "Win32_PerfRawData_HvStats_HyperVHypervisor" -Namespace "root\cimv2"

$logicalProcessorStatsCount = $logicalProcessorStats.LogicalProcessors

if($logicalProcessorStatsCount -lt 1)

{

  exit 

}

$totalLogicalProcessorStats=Get-CimInstance -ClassName "Win32_PerfRawData_HvStats_HyperVHypervisorLogicalProcessor" -Namespace "root\cimv2" -Filter "Name='_Total'"

ForEach ($item in $totalLogicalProcessorStats)

 {

  $percentTotalRunTime_First= $item.PercentTotalRunTime

  $timestamp_PerfTime_First = $item.Timestamp_PerfTime

 }

 

 Start-Sleep $timePeriodSeconds



 $totalLogicalProcessorStats=Get-CimInstance -ClassName "Win32_PerfRawData_HvStats_HyperVHypervisorLogicalProcessor" -Namespace "root\cimv2" -Filter "Name='_Total'"

ForEach ($item in $totalLogicalProcessorStats)

 {

  $frequency_PerfTime = $item.Frequency_PerfTime

  $percentTotalRunTime_Second= $item.PercentTotalRunTime

  $timestamp_PerfTime_Second = $item.Timestamp_PerfTime

 }



if($timestamp_PerfTime_Second - $timestamp_PerfTime_First -eq 0)

{

    $momApi.LogScriptEvent("HostCPUUtilizationProvider.ps1", 101, 0, "The readings were not taken properly and hence the script could not be executed.")

    exit 1

}



#Calculate CPU readings with formula 

$CPU = ( $percentTotalRunTime_Second - $percentTotalRunTime_First)/($timestamp_PerfTime_Second - $timestamp_PerfTime_First) 

$CPU = ($frequency_PerfTime * $CPU) / $logicalProcessorStatsCount

$CPU = [int]($CPU / 100000)



if( $CPU -gt 100)

{

 $CPU = 100

}

if( $CPU -lt 0) 

{

 $CPU = 0

}



$CPU = [int]($CPU)

$momBag.AddValue("Utilization", $CPU)                        

$momBag



</Script></ScriptBody>

    <ObjectName>Win32_PerfRawData_HvStats_HyperVHypervisorLogicalProcessor</ObjectName>

    <CounterName>PercentTotalRunTime</CounterName>

    <InstanceName>_Total</InstanceName>

    <Value>$Data/Property[@Name='Utilization']$</Value>

  </Configuration>

</UnitMonitor>