如何在AWS CDK中自动获取CloudWatch指标的所有维度？

CDK本身是静态基础设施即代码工具，在synth阶段生成CloudFormation模板，因此没有直接内置调用list-metrics的方法，但可以通过两种思路实现你的需求：

方案一：构建时通过AWS SDK获取维度（静态生成告警） #

在CDK代码执行阶段（本地synth或CI/CD部署时），直接调用AWS CloudWatch SDK的list-metrics接口，获取指定命名空间下的表名维度，再批量生成告警资源。

代码示例（TypeScript） #

import * as cdk from 'aws-cdk-lib';
import { Construct } from 'constructs';
import * as cloudwatch from 'aws-cdk-lib/aws-cloudwatch';
import { CloudWatchClient, ListMetricsCommand } from "@aws-sdk/client-cloudwatch";

export class TableMetricAlarmsStack extends cdk.Stack {
  constructor(scope: Construct, id: string, props?: cdk.StackProps) {
    super(scope, id, props);

    // 初始化CloudWatch客户端
    const cwClient = new CloudWatchClient({ region: this.region });

    // 调用list-metrics获取指定维度的指标
    const listMetricsCmd = new ListMetricsCommand({
      Namespace: "Your-Namespace", // 替换为你的指标命名空间
      MetricName: "Your-Metric-Name", // 替换为你的指标名
      Dimensions: [{ Name: "TableName" }] // 指定要过滤的维度名称
    });

    // 同步请求（synth阶段执行）
    const metricsResponse = await cwClient.send(listMetricsCmd);
    // 提取所有表名维度值
    const tableNames = metricsResponse.Metrics?.map(metric => 
      metric.Dimensions?.find(dim => dim.Name === "TableName")?.Value
    ).filter(Boolean) as string[];

    // 为每个表创建告警
    tableNames.forEach(tableName => {
      new cloudwatch.Alarm(this, `Table-Alarm-${tableName}`, {
        metric: new cloudwatch.Metric({
          namespace: "Your-Namespace",
          metricName: "Your-Metric-Name",
          dimensionsMap: { TableName: tableName },
          statistic: "Average",
          period: cdk.Duration.minutes(5),
        }),
        threshold: 100, // 替换为你的告警阈值
        evaluationPeriods: 2,
        comparisonOperator: cloudwatch.ComparisonOperator.GREATER_THAN_THRESHOLD,
        alarmDescription: `表${tableName}的指标超过阈值`,
        // 可添加告警动作，如SNS Topic
        // alarmActions: [new cloudwatch.SnsAction(yourSnsTopic)]
      });
    });
  }
}

注意事项 #

• 执行CDK synth/deploy的环境需要有cloudwatch:ListMetrics权限；
• 新表的指标需要先被CloudWatch采集到，才会出现在list-metrics结果中，新增表后需要重新运行CDK部署才能生成对应告警；
• 所有告警会被纳入CDK的基础设施管理，可通过cdk diff查看变更。

方案二：运行时动态创建告警（自动发现新表） #

如果希望新增表无需修改或重新部署CDK代码，可以用Lambda+EventBridge实现定期同步：Lambda调用list-metrics获取最新维度，动态创建/更新CloudWatch告警。

1. Lambda代码（Python） #

import boto3
import os

cw = boto3.client('cloudwatch')

def lambda_handler(event, context):
    namespace = os.environ['NAMESPACE']
    metric_name = os.environ['METRIC_NAME']
    threshold = float(os.environ['ALARM_THRESHOLD'])

    # 获取所有带TableName维度的指标
    response = cw.list_metrics(
        Namespace=namespace,
        MetricName=metric_name,
        Dimensions=[{'Name': 'TableName'}]
    )

    # 提取并去重表名
    table_names = set()
    for metric in response['Metrics']:
        for dim in metric['Dimensions']:
            if dim['Name'] == 'TableName':
                table_names.add(dim['Value'])

    # 为每个表创建或更新告警
    for table_name in table_names:
        alarm_name = f"Table-Alarm-{table_name}"
        cw.put_metric_alarm(
            AlarmName=alarm_name,
            AlarmDescription=f"表{table_name}的指标超过阈值",
            Namespace=namespace,
            MetricName=metric_name,
            Dimensions=[{'Name': 'TableName', 'Value': table_name}],
            Statistic='Average',
            Period=300, # 5分钟
            EvaluationPeriods=2,
            Threshold=threshold,
            ComparisonOperator='GreaterThanThreshold',
            # 替换为你的告警目标，如SNS Topic ARN
            # AlarmActions=['arn:aws:sns:us-east-1:123456789012:your-topic']
        )

    return {'statusCode': 200, 'body': f"已处理{len(table_names)}个表的告警配置"}

2. CDK代码定义Lambda和EventBridge规则 #

import * as cdk from 'aws-cdk-lib';
import { Construct } from 'constructs';
import * as lambda from 'aws-cdk-lib/aws-lambda';
import * as events from 'aws-cdk-lib/aws-events';
import * as targets from 'aws-cdk-lib/aws-events-targets';
import * as iam from 'aws-cdk-lib/aws-iam';

export class DynamicAlarmSyncStack extends cdk.Stack {
  constructor(scope: Construct, id: string, props?: cdk.StackProps) {
    super(scope, id, props);

    const alarmSyncLambda = new lambda.Function(this, 'AlarmSyncLambda', {
      runtime: lambda.Runtime.PYTHON_3_11,
      handler: 'index.lambda_handler',
      code: lambda.Code.fromAsset('lambda'), // 存放Lambda代码的目录
      environment: {
        NAMESPACE: "Your-Namespace",
        METRIC_NAME: "Your-Metric-Name",
        ALARM_THRESHOLD: "100"
      }
    });

    // 给Lambda授予必要权限
    alarmSyncLambda.addToRolePolicy(new iam.PolicyStatement({
      actions: ['cloudwatch:ListMetrics', 'cloudwatch:PutMetricAlarm'],
      resources: ['*']
    }));

    // 配置每天凌晨执行一次同步
    new events.Rule(this, 'DailyAlarmSyncRule', {
      schedule: events.Schedule.cron({ hour: '0', minute: '0' }),
      targets: [new targets.LambdaFunction(alarmSyncLambda)]
    });
  }
}

优缺点 #

• 优势：自动发现新表，无需重新部署CDK；告警配置随指标动态更新；
• 劣势：告警由Lambda动态创建，不在CDK的资源清单中，需要通过CloudWatch控制台或API查看管理。

内容的提问来源于stack exchange，提问作者Scooby