如何使用Java SDK创建Azure Monitor指标警报

1. 添加依赖 #

先在Maven项目中引入Azure Monitor Alerts SDK和身份验证相关依赖：

<dependencies>
    <!-- Azure Monitor Alerts SDK -->
    <dependency>
        <groupId>com.azure.resourcemanager</groupId>
        <artifactId>azure-resourcemanager-monitor</artifactId>
        <version>2.34.0</version>
    </dependency>
    <!-- Azure Identity 用于身份认证 -->
    <dependency>
        <groupId>com.azure</groupId>
        <artifactId>azure-identity</artifactId>
        <version>1.12.0</version>
    </dependency>
</dependencies>

2. 核心实现代码 #

以下是创建虚拟机CPU使用率超阈值警报的完整代码：

import com.azure.core.credential.DefaultAzureCredential;
import com.azure.resourcemanager.monitor.MonitorManager;
import com.azure.resourcemanager.monitor.models.*;

import java.time.Duration;
import java.util.Collections;

public class VmCpuAlertCreator {
    public static void main(String[] args) {
        // 替换为你的实际参数
        String subscriptionId = "your-subscription-id";
        String resourceGroupName = "your-vm-resource-group";
        String vmResourceId = "/subscriptions/your-subscription-id/resourceGroups/your-vm-resource-group/providers/Microsoft.Compute/virtualMachines/your-vm-name";
        String alertRuleName = "VM-CPU-Over-80-Alert";
        double cpuThreshold = 80.0;
        String vmRegion = "eastus"; // 替换为虚拟机所在区域

        // 初始化Monitor管理器
        MonitorManager monitorManager = MonitorManager.authenticate(new DefaultAzureCredential(), subscriptionId);

        // 定义CPU使用率的警报条件
        MetricAlertSingleCondition cpuCondition = new MetricAlertSingleCondition()
                .withOperator(ComparisonOperationType.GREATER_THAN)
                .withThreshold(cpuThreshold)
                .withTimeAggregation(TimeAggregationType.AVERAGE)
                .withMetricTrigger(new MetricTrigger()
                        .withMetricName("Percentage CPU")
                        .withNamespace("Microsoft.Compute/virtualMachines")
                        .withResourceLocation(vmRegion)
                        .withOperator(ComparisonOperationType.GREATER_THAN)
                        .withThreshold(cpuThreshold)
                        .withTimeAggregation(TimeAggregationType.AVERAGE)
                        .withTimeWindow(Duration.ofMinutes(5))
                        .withTimeAggregation(TimeAggregationType.AVERAGE));

        // 构建警报规则配置
        MetricAlertResource alertRule = new MetricAlertResource()
                .withDescription("虚拟机CPU使用率超过" + cpuThreshold + "%时触发警报")
                .withSeverity(AlertSeverityLevel.S2) // S0到S4，级别越高越紧急
                .withEnabled(true)
                .withScopes(Collections.singletonList(vmResourceId))
                .withEvaluationFrequency(Duration.ofMinutes(1)) // 每分钟检查一次
                .withWindowSize(Duration.ofMinutes(5)) // 基于过去5分钟的统计数据评估
                .withCriteria(new MetricAlertMultipleResourceMultipleMetricCriteria()
                        .withAllOf(Collections.singletonList(cpuThreshold)));

        // 如果需要绑定动作组（邮件/短信通知等），取消下面注释并替换动作组ID
        //.withActions(Collections.singletonList(new MetricAlertAction().withActionGroupId("/subscriptions/your-subscription-id/resourceGroups/your-rg/providers/microsoft.insights/actionGroups/your-action-group")));

        // 创建警报规则
        monitorManager.metricAlerts().define(alertRuleName)
                .withExistingResourceGroup(resourceGroupName)
                .withAlertRule(alertRule)
                .create();

        System.out.println("CPU警报规则创建完成");
    }
}

3. 关键配置说明 #

• 认证逻辑：DefaultAzureCredential会自动从环境变量、Azure CLI、托管身份等渠道获取凭据，适配开发和生产环境。
• 指标参数：MetricName固定为Percentage CPU，Namespace对应虚拟机资源的命名空间Microsoft.Compute/virtualMachines。
• 评估规则：timeWindow是统计指标的时间范围，evaluationFrequency是检查频率，可根据业务需求调整。
• 动作组：若需要触发通知，需提前在Azure门户创建动作组，再在代码中传入其资源ID。

注意事项 #

• 运行代码的身份（本地开发账户/生产环境托管身份）必须拥有Microsoft.Insights/metricAlerts/write权限。
• 务必替换代码中所有占位符为实际资源信息。

内容的提问来源于stack exchange，提问作者lpdevlop