无法查看Lambda函数的CloudWatch日志且执行超时，求调试方案

问题背景 #

用CDK创建AWS API Gateway并关联searchLambda，发送GET请求/?search=something&criteria=city时，请求会集成到searchLambda，理论上Lambda应接收参数并执行工作流，但调用API时出现超时错误，且Lambda顶部添加的console.log未在CloudWatch中输出。

错误日志 #

API Gateway日志：

(e276cbf8-8445-437b-93f7-fab1b32d8a22) Lambda execution failed with status 200 due to customer function error: 2023-03-01T19:59:11.265Z 9e793aa6-6e64-41e8-8237-b08ee11512be Task timed out after 10.80 seconds. Lambda request id: 9e793aa6-6e64-41e8-8237-b08ee11512be

注：已将Lambda超时时间从3秒调至10秒，问题仍存在

Lambda CloudWatch日志：

2023-03-01T16:44:29.503-05:00
INIT_START Runtime Version: nodejs:18.v5    Runtime Version ARN: arn:aws:lambda:us-east-1::runtime:c869d752e4ae21a3945cfcb3c1ff2beb1f160d7bcec3b0a8ef7caceae73c055f
INIT_START Runtime Version: nodejs:18.v5 Runtime Version ARN: arn:aws:lambda:us-east-1::runtime:c869d752e4ae21a3945cfcb3c1ff2beb1f160d7bcec3b0a8ef7caceae73c055f

2023-03-01T16:44:41.491-05:00
START RequestId: 6c70644b-6ceb-4b9a-83f3-b90b36aa6ebd Version: $LATEST
START RequestId: 6c70644b-6ceb-4b9a-83f3-b90b36aa6ebd Version: $LATEST

2023-03-01T16:44:41.652-05:00
2023-03-01T21:44:41.652Z 6c70644b-6ceb-4b9a-83f3-b90b36aa6ebd Task timed out after 10.21 seconds

2023-03-01T21:44:41.652Z 6c70644b-6ceb-4b9a-83f3-b90b36aa6ebd Task timed out after 10.21 seconds

2023-03-01T16:44:41.652-05:00
END RequestId: 6c70644b-6ceb-4b9a-83f3-b90b36aa6ebd
END RequestId: 6c70644b-6ceb-4b9a-83f3-b90b36aa6ebd

2023-03-01T16:44:41.652-05:00
REPORT RequestId: 6c70644b-6ceb-4b9a-83f3-b90b36aa6ebd  Duration: 10208.91 ms   Billed Duration: 10000 ms   Memory Size: 128 MB Max Memory Used: 128 MB 

Lambda配置 #

{
    "FunctionName": "SearchFlowStack-searchLambda02D5D642-rXVh6Hop7PM3",
    "FunctionArn": "arn:aws:lambda:us-east-1:201682123230:function:SearchFlowStack-searchLambda02D5D642-rXVh6Hop7PM3",
    "Runtime": "nodejs18.x",
    "Role": "arn:aws:iam::201682123230:role/SearchFlowStack-searchLambdaServiceRole4ED34712-1HDDNCA114A09",
    "Handler": "index.handler",
    "CodeSize": 5479738,
    "Description": "",
    "Timeout": 10,
    "MemorySize": 128,
    "LastModified": "2023-03-01T19:57:32.000+0000",
    "CodeSha256": "HCyIEDuPx69arYaP+yKdkfRDmXLrVuiy/BO+AsfteDA=",
    "Version": "$LATEST",
    "VpcConfig": {
        "SubnetIds": [],
        "SecurityGroupIds": [],
        "VpcId": ""
    },
    "Environment": {
        "Variables": {
            "STATE_MACHINE_ARN": "arn:aws:states:us-east-1:201682123230:stateMachine:cfnStepFunction-Iz9cXZoGzegj",
            "AWS_NODEJS_CONNECTION_REUSE_ENABLED": "1"
        }
    },
    "TracingConfig": {
        "Mode": "PassThrough"
    },
    "RevisionId": "83cb39a5-4e5e-48ca-a838-cf28344b2e04",
    "State": "Active",
    "LastUpdateStatus": "Successful",
    "PackageType": "Zip",
    "Architectures": [
        "x86_64"
    ]
}

Lambda代码片段 #

import { Lambda, StepFunctions } from 'aws-sdk';
import { APIGatewayProxyEvent } from 'aws-lambda';
import { SearchFlowStack } from '../../infra/SearchFlow';

export const handler = async (event: APIGatewayProxyEvent): Promise<any> => {
  console.log(`#### Step00. Started!!`)
  const { search, criteria } = event.queryStringParameters || {};

  const stepFunctions = new StepFunctions();
  const lambda = new Lambda();
  const { cfnStepFunction, cfnSearchResultsFunction } = SearchFlowStack;

  // Start the Step Function with the given criteria
  const params = {
    stateMachineArn: cfnStepFunction.attrArn,
    input: JSON.stringify({ searchTerm: search, criteria })
  };
.... ....

调试思路与方法 #

• 排查代码初始化阻塞：Lambda未输出初始console.log，说明代码可能在handler执行前就卡住。检查导入的SearchFlowStack是否存在同步阻塞操作，比如静态初始化时执行了耗时的CDK资源解析或网络请求。
• 提升内存配置：当前Lambda内存为128MB（最低配置），Lambda内存与CPU性能成正比，内存不足会导致CPU受限。尝试将内存提升至256MB或更高，观察是否仍超时。
• 检查IAM权限：确认Lambda角色拥有调用Step Functions的states:StartExecution权限，若权限缺失，SDK调用会静默等待直到Lambda超时。
• 添加SDK调用超时：为StepFunctions和Lambda客户端设置超时时间，避免无限等待。示例：

  const stepFunctions = new StepFunctions({
    httpOptions: { timeout: 5000 } // 设置5秒超时
  });
  

• 本地模拟测试：使用aws-sdk-mock或serverless-offline在本地模拟API Gateway请求，检查代码是否能正常执行，定位初始化或StepFunction调用环节的问题。
• 检查CloudWatch日志权限：确认Lambda角色拥有logs:CreateLogGroup、logs:CreateLogStream、logs:PutLogEvents权限，否则日志无法写入CloudWatch。
• 排查StepFunction状态：检查目标StepFunction是否正常（如是否处于暂停状态、资源是否存在），若目标资源不可用，StartExecution调用会一直等待直到Lambda超时。

内容的提问来源于stack exchange，提问作者track_p54skate