如何将Azure VM告警规则脚本改造为跨订阅/资源组的自动化脚本

问题背景 #

现有PowerShell脚本可创建Azure VM的监控告警规则，但每次适配不同订阅、资源组的VM时，需手动修改告警名称、目标资源范围等信息，无法批量自动化处理。

改造步骤 #

• 提取可配置参数：将订阅ID、动作组ID、告警规则名称前缀、目标资源组（可选）、告警阈值等硬编码值，改为脚本参数，运行时可灵活传入。比如用param块定义参数，支持指定单个/多个订阅、资源组，或覆盖默认阈值。
• 动态获取目标VM：使用Get-AzVM cmdlet自动拉取指定范围内的所有VM（可指定资源组，或全订阅），无需手动逐个指定VM名称和资源路径。
• 重构告警规则配置：把每个告警的指标名称、阈值、窗口大小、频率、描述等信息整理成一个配置数组，循环遍历每个VM和每个告警配置，自动拼接告警规则名称、生成目标资源范围路径。
• 增加幂等性检查：在创建告警规则前，先通过Get-AzMetricAlertRuleV2检查同名规则是否已存在，避免重复创建导致报错。
• 支持多订阅切换：如果需要适配多个订阅，添加订阅遍历逻辑，先切换到目标订阅，再执行VM获取和告警创建操作。

改造后的示例脚本 #

param(
    [Parameter(Mandatory=$true)]
    [string[]]$SubscriptionIds,
    [Parameter(Mandatory=$true)]
    [string]$ActionGroupResourceId,
    [string]$AlertRulePrefix = "ClientName - Virtual machines",
    [string]$TargetResourceGroupName, # 留空则处理订阅下所有资源组的VM
    [int]$CpuThreshold = 96,
    [long]$MemoryThreshold = 1000000000,
    [int]$DiskIopsThreshold = 95,
    [int]$VmAvailabilityThreshold = 1
)

# 定义告警规则配置数组
$alertConfigs = @(
    @{
        MetricName = "Percentage CPU"
        Operator = "GreaterThan"
        Threshold = $CpuThreshold
        WindowSize = "0:30"
        Frequency = "0:15"
        Description = "Action will be triggered when CPU usage is greater than threshold"
        Severity = 0
    },
    @{
        MetricName = "Available Memory Bytes"
        Operator = "LessThan"
        Threshold = $MemoryThreshold
        WindowSize = "0:30"
        Frequency = "0:30"
        Description = "Action will be triggered when Available Memory Bytes is less than threshold"
        Severity = 0
    },
    @{
        MetricName = "Data Disk IOPS Consumed Percentage"
        Operator = "GreaterThan"
        Threshold = $DiskIopsThreshold
        WindowSize = "1:0"
        Frequency = "0:30"
        Description = "Action will be triggered when Data Disk IOPS usage is greater than threshold"
        Severity = 0
    },
    @{
        MetricName = "OS Disk IOPS Consumed Percentage"
        Operator = "GreaterThan"
        Threshold = $DiskIopsThreshold
        WindowSize = "1:0"
        Frequency = "0:30"
        Description = "Action will be triggered when OS Disk IOPS usage is greater than threshold"
        Severity = 0
    },
    @{
        MetricName = "VmAvailabilityMetric"
        Operator = "LessThan"
        Threshold = $VmAvailabilityThreshold
        WindowSize = "0:15"
        Frequency = "0:15"
        Description = "Action will be triggered when VM Availability is less than threshold."
        Severity = 0
    }
)

# 初始化动作组对象
$actionGroup = [Microsoft.Azure.Management.Monitor.Management.Models.ActivityLogAlertActionGroup]::New($ActionGroupResourceId)

# 遍历每个订阅
foreach ($subId in $SubscriptionIds) {
    # 切换到目标订阅
    Set-AzContext -Subscription $subId | Out-Null
    Write-Host "Processing subscription: $subId"

    # 获取目标VM
    $vms = if ($TargetResourceGroupName) {
        Get-AzVM -ResourceGroupName $TargetResourceGroupName
    } else {
        Get-AzVM
    }

    if (-not $vms) {
        Write-Host "No VMs found in subscription $subId"
        continue
    }

    # 遍历每个VM
    foreach ($vm in $vms) {
        $vmName = $vm.Name
        $vmResourceGroup = $vm.ResourceGroupName
        $vmRegion = $vm.Location
        $vmResourceId = $vm.Id
        Write-Host "Processing VM: $vmName in resource group $vmResourceGroup"

        # 遍历每个告警配置
        foreach ($config in $alertConfigs) {
            # 生成告警规则名称
            $alertRuleName = "$AlertRulePrefix - $vmName - $($config.MetricName) - Critical"
            
            # 检查告警规则是否已存在
            $existingRule = Get-AzMetricAlertRuleV2 -ResourceGroupName $vmResourceGroup -Name $alertRuleName -ErrorAction SilentlyContinue
            if ($existingRule) {
                Write-Host "Alert rule '$alertRuleName' already exists, skipping..."
                continue
            }

            # 创建告警条件
            $criteria = New-AzMetricAlertRuleV2Criteria `
                -MetricName $config.MetricName `
                -MetricNameSpace "Microsoft.Compute/virtualMachines" `
                -TimeAggregation Average `
                -Operator $config.Operator `
                -Threshold $config.Threshold

            # 创建告警规则
            try {
                Add-AzMetricAlertRuleV2 `
                    -Name $alertRuleName `
                    -ResourceGroupName $vmResourceGroup `
                    -WindowSize $config.WindowSize `
                    -Frequency $config.Frequency `
                    -TargetResourceScope $vmResourceId `
                    -TargetResourceType "Microsoft.Compute/virtualMachines" `
                    -TargetResourceRegion $vmRegion `
                    -Description $config.Description `
                    -Severity $config.Severity `
                    -ActionGroup $actionGroup `
                    -Condition $criteria
                Write-Host "Successfully created alert rule: $alertRuleName"
            } catch {
                Write-Error "Failed to create alert rule '$alertRuleName': $_"
            }
        }
    }
}

使用说明 #

1. 运行脚本时，传入必填参数：.\Automate-VMAlerts.ps1 -SubscriptionIds "sub-id-1","sub-id-2" -ActionGroupResourceId "/subscriptions/xxx/resourceGroups/alert/providers/Microsoft.Insights/actiongroups/alertaction"
2. 可选参数：指定-TargetResourceGroupName处理单个资源组的VM，或调整阈值参数覆盖默认值。

内容的提问来源于stack exchange，提问作者Md Nizam Uddin