如何使用AWS CDK获取现有EC2实例ID存入数组并配置CloudWatch告警

CDK目前没有内置类似vpc.fromLookup的批量查询所有EC2实例的方法，你可以通过以下两种方案实现实例列表拉取，再批量配置CloudWatch CPU告警和SNS通知：

方案1：CDK运行时调用AWS SDK拉取实例列表 #

直接在CDK栈初始化前调用EC2 SDK的describeInstances接口拉取实例ID，再遍历创建告警规则，示例代码（TypeScript环境）如下：

1. 先安装依赖：

npm install @aws-sdk/client-ec2

2. 栈代码示例：

import { EC2Client, DescribeInstancesCommand } from "@aws-sdk/client-ec2";
import { App, Stack, StackProps, Duration } from 'aws-cdk-lib';
import { Metric, Alarm, ComparisonOperator } from 'aws-cdk-lib/aws-cloudwatch';
import { SnsAction } from 'aws-cdk-lib/aws-cloudwatch-actions';
import { Topic } from 'aws-cdk-lib/aws-sns';

const app = new App();
class Ec2CpuMonitorStack extends Stack {
  constructor(scope: App, id: string, instanceIds: string[], props?: StackProps) {
    super(scope, id, props);

    // 替换为你的SNS通知主题ARN
    const alarmNotifyTopic = Topic.fromTopicArn(this, 'AlarmNotifyTopic', 'arn:aws-cn:sns:cn-north-1:123456789012:ec2-alarm-topic');

    // 遍历为每个实例创建CPU利用率告警
    instanceIds.forEach(instanceId => {
      const cpuUtilMetric = new Metric({
        namespace: 'AWS/EC2',
        metricName: 'CPUUtilization',
        dimensionsMap: { InstanceId: instanceId },
        period: Duration.minutes(5),
        statistic: 'Average'
      });

      new Alarm(this, `CpuHighAlarm-${instanceId}`, {
        metric: cpuUtilMetric,
        threshold: 80, // 可自定义阈值，示例为80%
        evaluationPeriods: 3,
        comparisonOperator: ComparisonOperator.GREATER_THAN_THRESHOLD,
        alarmDescription: `EC2实例 ${instanceId} 连续3个周期平均CPU利用率超过80%`
      }).addAlarmAction(new SnsAction(alarmNotifyTopic));
    });
  }
}

// 拉取当前区域所有EC2实例ID
const getAllEc2InstanceIds = async () => {
  const ec2Client = new EC2Client({ region: process.env.CDK_DEFAULT_REGION });
  const resp = await ec2Client.send(new DescribeInstancesCommand({
    // 可按需添加过滤器，比如只拉取运行中实例、按标签过滤等
    Filters: [{ Name: 'instance-state-name', Values: ['running'] }]
  }));
  const instanceIds: string[] = [];
  resp.Reservations?.forEach(reservation => {
    reservation.Instances?.forEach(instance => {
      if (instance.InstanceId) instanceIds.push(instance.InstanceId);
    });
  });
  return instanceIds;
};

// 初始化栈并执行synth
getAllEc2InstanceIds().then(instanceIds => {
  new Ec2CpuMonitorStack(app, 'Ec2CpuMonitorStack', instanceIds);
  app.synth();
});

注意事项 #

• 执行cdk synth/cdk deploy的环境需要具备对应AWS账号的ec2:DescribeInstances权限
• 示例默认拉取所有运行中实例，你可以修改DescribeInstancesCommand的Filters参数按需过滤实例

方案2：预拉取实例ID存入CDK Context #

如果你不想在CDK运行时调用SDK，可以提前导出实例ID写入CDK上下文：

1. 本地执行AWS CLI命令导出运行中实例ID：

aws ec2 describe-instances --filters "Name=instance-state-name,Values=running" --query "Reservations[*].Instances[*].InstanceId" --output text | tr '\t' ',' > instance_ids.txt

2. 在执行cdk命令时通过context传入实例ID：

cdk deploy --context instanceIds=$(cat instance_ids.txt)

3. 在栈中读取context参数处理：

const instanceIds = (this.node.tryGetContext('instanceIds') as string).split(',');
// 后续遍历创建告警逻辑和方案1一致

提示：两种方案都是部署时静态拉取实例列表，如果后续账号下新增EC2实例，需要重新执行cdk deploy才能为新实例配置告警。如果需要自动适配新增实例，可以搭配EventBridge监听EC2实例创建事件，触发Lambda动态创建告警规则。

内容的提问来源于stack exchange，提问作者FilipKamenar