如何获取RDS集群实例最大连接数，配置CloudWatch告警阈值为其75%？

针对你遇到的问题，这里提供三种实用的解决思路，适配不同场景需求：

方法一：预定义实例类型与最大连接数映射表 #

Aurora MySQL的默认max_connections值与实例类型、引擎版本强相关，官方文档有明确对应关系。你可以预先整理一个映射表，直接在代码中根据实例类型获取最大连接数，再计算75%作为阈值。

代码示例 #

// 预定义对应引擎版本（2.09.2）的实例类型与最大连接数映射
const auroraMysqlMaxConnections = {
  't3.medium': 1000, // 请根据官方文档确认对应版本的实际值
  't3.large': 2000,
  // 按需添加其他需要的实例类型
};

// 定义实例类型（可复用）
const instanceType = ec2.InstanceType.of(ec2.InstanceClass.T3, ec2.InstanceSize.MEDIUM);

// 创建RDS集群
const TestCluster = new rds.DatabaseCluster(this, "TestDbCluster", {
  engine: rds.DatabaseClusterEngine.auroraMysql({
    version: rds.AuroraMysqlEngineVersion.VER_2_09_2,
  }),
  instances: 1,
  instanceProps: {
    instanceType: instanceType,
    vpcSubnets: {
      subnetType: ec2.SubnetType.PRIVATE_ISOLATED,
    },
    vpc: vpc,
  },
  clusterIdentifier: "TestDbCluster",
});

// 计算告警阈值
const maxConnections = auroraMysqlMaxConnections[instanceType.toString()];
const threshold = maxConnections * 0.75;

// 创建CloudWatch告警
const ConnectionsMetric = TestCluster.metricDatabaseConnections();
const dbConnectionsAlarm = ConnectionsMetric.createAlarm(this, "TestAlarm", {
  alarmName: "DB-DbConnections-Alarm",
  threshold: threshold,
  evaluationPeriods: 1,
});

优缺点 #

• 优点：实现简单，无额外资源开销
• 缺点：需要手动维护映射表，引擎版本升级或实例类型变更时需同步更新

方法二：通过自定义资源动态查询集群参数 #

利用CDK的自定义资源（Custom Resource）调用AWS SDK查询RDS集群的max_connections参数值，动态计算阈值，适配实例类型和参数的变更。

代码示例 #

import { CustomResource, CustomResourceProvider } from 'aws-cdk-lib/custom-resources';
import { PolicyStatement } from 'aws-cdk-lib/aws-iam';
import { Fn } from 'aws-cdk-lib';

// 创建RDS集群
const TestCluster = new rds.DatabaseCluster(this, "TestDbCluster", {
  engine: rds.DatabaseClusterEngine.auroraMysql({
    version: rds.AuroraMysqlEngineVersion.VER_2_09_2,
  }),
  instances: 1,
  instanceProps: {
    instanceType: ec2.InstanceType.of(ec2.InstanceClass.T3, ec2.InstanceSize.MEDIUM),
    vpcSubnets: {
      subnetType: ec2.SubnetType.PRIVATE_ISOLATED,
    },
    vpc: vpc,
  },
  clusterIdentifier: "TestDbCluster",
});

// 自定义资源：查询集群的max_connections参数
const maxConnectionsResource = new CustomResource(this, 'MaxConnectionsResource', {
  serviceToken: CustomResourceProvider.getOrCreate(this, 'Custom::RDSMaxConnections', {
    policyStatements: [
      new PolicyStatement({
        actions: ['rds:DescribeDBClusterParameters'],
        resources: [TestCluster.clusterArn],
      })
    ],
    onCreate: {
      service: 'RDS',
      action: 'describeDBClusterParameters',
      parameters: {
        DBClusterParameterGroupName: TestCluster.clusterParameterGroup!.parameterGroupName,
        Filters: [{ Name: 'parameter-name', Values: ['max_connections'] }]
      },
      physicalResourceId: { id: Date.now().toString() },
    },
    onUpdate: {
      service: 'RDS',
      action: 'describeDBClusterParameters',
      parameters: {
        DBClusterParameterGroupName: TestCluster.clusterParameterGroup!.parameterGroupName,
        Filters: [{ Name: 'parameter-name', Values: ['max_connections'] }]
      },
      physicalResourceId: { id: Date.now().toString() },
    },
  }),
});

// 从自定义资源中提取参数值，计算75%阈值
const maxConnections = maxConnectionsResource.getAttString('Parameters.0.ParameterValue');
const threshold = Fn.toNumber(maxConnections) * 0.75;

// 创建CloudWatch告警
const ConnectionsMetric = TestCluster.metricDatabaseConnections();
const dbConnectionsAlarm = ConnectionsMetric.createAlarm(this, "TestAlarm", {
  alarmName: "DB-DbConnections-Alarm",
  threshold: threshold,
  evaluationPeriods: 1,
});

优缺点 #

• 优点：动态获取参数值，无需手动维护映射，适配实例类型和参数变更
• 缺点：引入自定义资源，增加少量部署复杂度（会创建一个Lambda函数）

方法三：使用CloudWatch Metric Math实现比例告警 #

通过定期推送max_connections到CloudWatch自定义指标，再用Metric Math计算当前连接数占最大连接数的比例，直接对比例阈值（75%）告警。

步骤说明 #

1. 编写Lambda函数：定期调用RDS API查询max_connections，并推送到CloudWatch自定义指标
2. 在CDK中创建Metric Math告警，对比当前连接数与最大连接数的比例

告警代码示例 #

import * as cloudwatch from 'aws-cdk-lib/aws-cloudwatch';

const ConnectionsMetric = TestCluster.metricDatabaseConnections();
// 自定义指标：预先通过Lambda推送的max_connections值
const maxConnectionsMetric = new cloudwatch.Metric({
  namespace: 'Custom/RDS',
  metricName: 'MaxConnections',
  dimensionsMap: { DBClusterIdentifier: TestCluster.clusterIdentifier },
});

// 创建比例告警：当连接数占比超过75%时触发
const dbConnectionsAlarm = new cloudwatch.Alarm(this, "TestAlarm", {
  alarmName: "DB-DbConnections-Ratio-Alarm",
  evaluationPeriods: 1,
  metric: ConnectionsMetric.divide(maxConnectionsMetric).multiply(100), // 计算百分比
  threshold: 75,
  comparisonOperator: cloudwatch.ComparisonOperator.GREATER_THAN_THRESHOLD,
});

优缺点 #

• 优点：基于运行时实际参数值，告警更精准
• 缺点：需要额外配置Lambda和CloudWatch指标推送，增加运维成本

内容的提问来源于stack exchange，提问作者ShanWave007