Back to Details

serverless-architecture

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Serverless Architecture

无服务器架构

Overview

概述

Serverless architecture enables building complete applications without managing servers. Design event-driven, scalable systems using managed compute services, databases, and messaging systems. Pay only for actual usage with automatic scaling.
无服务器架构支持在无需管理服务器的情况下构建完整应用程序。使用托管计算服务、数据库和消息传递系统设计事件驱动的可扩展系统。仅按实际使用量付费，且支持自动扩缩容。

When to Use

适用场景

  • Event-driven applications
  • API backends and microservices
  • Real-time data processing
  • Batch jobs and scheduled tasks
  • Workflow automation
  • IoT data pipelines
  • Multi-tenant SaaS applications
  • Mobile app backends
  • 事件驱动型应用程序
  • API后端与微服务
  • 实时数据处理
  • 批处理作业与定时任务
  • 工作流自动化
  • IoT数据管道
  • 多租户SaaS应用程序
  • 移动应用后端

Implementation Examples

实现示例

1. Serverless Application Architecture

1. 无服务器应用架构

yaml
undefined
yaml
undefined

serverless.yml - Serverless Framework

serverless.yml - Serverless Framework

service: my-app
frameworkVersion: '3'
provider: name: aws runtime: nodejs18.x region: us-east-1 stage: ${opt:stage, 'dev'} memorySize: 256 timeout: 30 environment: STAGE: ${self:provider.stage} DYNAMODB_TABLE: ${self:service}-users-${self:provider.stage} SNS_TOPIC_ARN: arn:aws:sns:${self:provider.region}:${aws:accountId}:my-topic httpApi: cors: true iam: role: statements: - Effect: Allow Action: - dynamodb:Query - dynamodb:Scan - dynamodb:GetItem - dynamodb:PutItem - dynamodb:UpdateItem - dynamodb:DeleteItem Resource: "arn:aws:dynamodb:${self:provider.region}:${aws:accountId}:table/${self:provider.environment.DYNAMODB_TABLE}" - Effect: Allow Action: - sns:Publish Resource: ${self:provider.environment.SNS_TOPIC_ARN}
functions:

HTTP API endpoints

getUser: handler: src/handlers/getUser.handler events: - httpApi: path: /api/users/{id} method: GET
listUsers: handler: src/handlers/listUsers.handler events: - httpApi: path: /api/users method: GET
createUser: handler: src/handlers/createUser.handler events: - httpApi: path: /api/users method: POST

Event-driven functions

processUserCreated: handler: src/handlers/processUserCreated.handler events: - sns: arn: arn:aws:sns:${self:provider.region}:${aws:accountId}:user-created topicName: user-created
processPendingOrders: handler: src/handlers/processPendingOrders.handler timeout: 300 events: - schedule: rate: cron(0 2 * * ? *) enabled: true

S3 event handler

processImageUpload: handler: src/handlers/processImageUpload.handler events: - s3: bucket: my-uploads-${self:provider.stage} event: s3:ObjectCreated:* rules: - prefix: uploads/ - suffix: .jpg

SQS queue processor

processQueue: handler: src/handlers/processQueue.handler events: - sqs: arn: arn:aws:sqs:${self:provider.region}:${aws:accountId}:my-queue batchSize: 10 batchWindow: 5
resources: Resources: UsersTable: Type: AWS::DynamoDB::Table Properties: TableName: ${self:provider.environment.DYNAMODB_TABLE} AttributeDefinitions: - AttributeName: id AttributeType: S - AttributeName: createdAt AttributeType: N KeySchema: - AttributeName: id KeyType: HASH - AttributeName: createdAt KeyType: RANGE BillingMode: PAY_PER_REQUEST StreamSpecification: StreamViewType: NEW_AND_OLD_IMAGES
UserNotificationTopic:
  Type: AWS::SNS::Topic
  Properties:
    TopicName: user-created-${self:provider.stage}

ProcessingQueue:
  Type: AWS::SQS::Queue
  Properties:
    QueueName: my-queue-${self:provider.stage}
    VisibilityTimeout: 300
    MessageRetentionPeriod: 1209600
plugins:
  • serverless-python-requirements
  • serverless-plugin-tracing
  • serverless-offline
  • serverless-dynamodb-local
undefined
service: my-app
frameworkVersion: '3'
provider: name: aws runtime: nodejs18.x region: us-east-1 stage: ${opt:stage, 'dev'} memorySize: 256 timeout: 30 environment: STAGE: ${self:provider.stage} DYNAMODB_TABLE: ${self:service}-users-${self:provider.stage} SNS_TOPIC_ARN: arn:aws:sns:${self:provider.region}:${aws:accountId}:my-topic httpApi: cors: true iam: role: statements: - Effect: Allow Action: - dynamodb:Query - dynamodb:Scan - dynamodb:GetItem - dynamodb:PutItem - dynamodb:UpdateItem - dynamodb:DeleteItem Resource: "arn:aws:dynamodb:${self:provider.region}:${aws:accountId}:table/${self:provider.environment.DYNAMODB_TABLE}" - Effect: Allow Action: - sns:Publish Resource: ${self:provider.environment.SNS_TOPIC_ARN}
functions:

HTTP API endpoints

getUser: handler: src/handlers/getUser.handler events: - httpApi: path: /api/users/{id} method: GET
listUsers: handler: src/handlers/listUsers.handler events: - httpApi: path: /api/users method: GET
createUser: handler: src/handlers/createUser.handler events: - httpApi: path: /api/users method: POST

Event-driven functions

processUserCreated: handler: src/handlers/processUserCreated.handler events: - sns: arn: arn:aws:sns:${self:provider.region}:${aws:accountId}:user-created topicName: user-created
processPendingOrders: handler: src/handlers/processPendingOrders.handler timeout: 300 events: - schedule: rate: cron(0 2 * * ? *) enabled: true

S3 event handler

processImageUpload: handler: src/handlers/processImageUpload.handler events: - s3: bucket: my-uploads-${self:provider.stage} event: s3:ObjectCreated:* rules: - prefix: uploads/ - suffix: .jpg

SQS queue processor

processQueue: handler: src/handlers/processQueue.handler events: - sqs: arn: arn:aws:sqs:${self:provider.region}:${aws:accountId}:my-queue batchSize: 10 batchWindow: 5
resources: Resources: UsersTable: Type: AWS::DynamoDB::Table Properties: TableName: ${self:provider.environment.DYNAMODB_TABLE} AttributeDefinitions: - AttributeName: id AttributeType: S - AttributeName: createdAt AttributeType: N KeySchema: - AttributeName: id KeyType: HASH - AttributeName: createdAt KeyType: RANGE BillingMode: PAY_PER_REQUEST StreamSpecification: StreamViewType: NEW_AND_OLD_IMAGES
UserNotificationTopic:
  Type: AWS::SNS::Topic
  Properties:
    TopicName: user-created-${self:provider.stage}

ProcessingQueue:
  Type: AWS::SQS::Queue
  Properties:
    QueueName: my-queue-${self:provider.stage}
    VisibilityTimeout: 300
    MessageRetentionPeriod: 1209600
plugins:
  • serverless-python-requirements
  • serverless-plugin-tracing
  • serverless-offline
  • serverless-dynamodb-local
undefined

2. Event-Driven Lambda Handler Pattern

2. 事件驱动Lambda处理器模式

javascript
// src/handlers/processUserCreated.js
const AWS = require('aws-sdk');
const dynamodb = new AWS.DynamoDB.DocumentClient();

const userService = require('../services/userService');
const emailService = require('../services/emailService');

exports.handler = async (event, context) => {
  console.log('Processing user created event:', JSON.stringify(event));

  try {
    // Parse SNS message
    const records = event.Records;

    for (const record of records) {
      const message = JSON.parse(record.Sns.Message);
      const userId = message.userId;

      // Get user details
      const user = await userService.getUser(userId);

      // Send welcome email
      await emailService.sendWelcomeEmail(user);

      // Initialize user preferences
      await dynamodb.put({
        TableName: process.env.DYNAMODB_TABLE,
        Item: {
          id: userId,
          preferences: {
            newsletter: true,
            notifications: true
          },
          createdAt: Date.now()
        }
      }).promise();

      // Log success
      console.log(`Successfully processed user creation for ${userId}`);
    }

    return {
      statusCode: 200,
      body: JSON.stringify({ message: 'Processed' })
    };
  } catch (error) {
    console.error('Error processing event:', error);
    throw error; // SNS will retry
  }
};

// src/handlers/processImageUpload.js
const AWS = require('aws-sdk');
const s3 = new AWS.S3();
const rekognition = new AWS.Rekognition();

exports.handler = async (event, context) => {
  try {
    for (const record of event.Records) {
      const bucket = record.s3.bucket.name;
      const key = record.s3.object.key;

      console.log(`Processing image: s3://${bucket}/${key}`);

      // Analyze image with Rekognition
      const labels = await rekognition.detectLabels({
        Image: {
          S3Object: {
            Bucket: bucket,
            Name: key
          }
        },
        MaxLabels: 10,
        MinConfidence: 70
      }).promise();

      // Create thumbnail
      await createThumbnail(bucket, key);

      // Index metadata
      await indexMetadata(bucket, key, labels);

      console.log(`Completed processing ${key}`);
    }
  } catch (error) {
    console.error('Error processing S3 event:', error);
    throw error;
  }
};

async function createThumbnail(bucket, key) {
  // Implementation
  return true;
}

async function indexMetadata(bucket, key, labels) {
  // Implementation
  return true;
}
javascript
// src/handlers/processUserCreated.js
const AWS = require('aws-sdk');
const dynamodb = new AWS.DynamoDB.DocumentClient();

const userService = require('../services/userService');
const emailService = require('../services/emailService');

exports.handler = async (event, context) => {
  console.log('Processing user created event:', JSON.stringify(event));

  try {
    // Parse SNS message
    const records = event.Records;

    for (const record of records) {
      const message = JSON.parse(record.Sns.Message);
      const userId = message.userId;

      // Get user details
      const user = await userService.getUser(userId);

      // Send welcome email
      await emailService.sendWelcomeEmail(user);

      // Initialize user preferences
      await dynamodb.put({
        TableName: process.env.DYNAMODB_TABLE,
        Item: {
          id: userId,
          preferences: {
            newsletter: true,
            notifications: true
          },
          createdAt: Date.now()
        }
      }).promise();

      // Log success
      console.log(`Successfully processed user creation for ${userId}`);
    }

    return {
      statusCode: 200,
      body: JSON.stringify({ message: 'Processed' })
    };
  } catch (error) {
    console.error('Error processing event:', error);
    throw error; // SNS will retry
  }
};

// src/handlers/processImageUpload.js
const AWS = require('aws-sdk');
const s3 = new AWS.S3();
const rekognition = new AWS.Rekognition();

exports.handler = async (event, context) => {
  try {
    for (const record of event.Records) {
      const bucket = record.s3.bucket.name;
      const key = record.s3.object.key;

      console.log(`Processing image: s3://${bucket}/${key}`);

      // Analyze image with Rekognition
      const labels = await rekognition.detectLabels({
        Image: {
          S3Object: {
            Bucket: bucket,
            Name: key
          }
        },
        MaxLabels: 10,
        MinConfidence: 70
      }).promise();

      // Create thumbnail
      await createThumbnail(bucket, key);

      // Index metadata
      await indexMetadata(bucket, key, labels);

      console.log(`Completed processing ${key}`);
    }
  } catch (error) {
    console.error('Error processing S3 event:', error);
    throw error;
  }
};

async function createThumbnail(bucket, key) {
  // Implementation
  return true;
}

async function indexMetadata(bucket, key, labels) {
  // Implementation
  return true;
}

3. Orchestration with Step Functions

3. 使用Step Functions进行编排

json
{
  "Comment": "Order processing workflow",
  "StartAt": "ValidateOrder",
  "States": {
    "ValidateOrder": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:validateOrder",
      "Next": "CheckInventory",
      "Catch": [
        {
          "ErrorEquals": ["InvalidOrder"],
          "Next": "OrderFailed"
        }
      ]
    },
    "CheckInventory": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:checkInventory",
      "Next": "InventoryDecision"
    },
    "InventoryDecision": {
      "Type": "Choice",
      "Choices": [
        {
          "Variable": "$.inStock",
          "BooleanEquals": true,
          "Next": "ProcessPayment"
        }
      ],
      "Default": "OutOfStock"
    },
    "ProcessPayment": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:processPayment",
      "Next": "PaymentDecision",
      "Retry": [
        {
          "ErrorEquals": ["PaymentError"],
          "IntervalSeconds": 2,
          "MaxAttempts": 3,
          "BackoffRate": 2.0
        }
      ]
    },
    "PaymentDecision": {
      "Type": "Choice",
      "Choices": [
        {
          "Variable": "$.paymentApproved",
          "BooleanEquals": true,
          "Next": "ShipOrder"
        }
      ],
      "Default": "PaymentFailed"
    },
    "ShipOrder": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:shipOrder",
      "Next": "NotifyCustomer"
    },
    "NotifyCustomer": {
      "Type": "Task",
      "Resource": "arn:aws:states:::sns:publish",
      "Parameters": {
        "TopicArn": "arn:aws:sns:region:account:order-updates",
        "Message": {
          "orderId.$": "$.orderId",
          "status": "shipped"
        }
      },
      "Next": "OrderSuccess"
    },
    "OrderSuccess": {
      "Type": "Succeed"
    },
    "OutOfStock": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:notifyOutOfStock",
      "Next": "OrderFailed"
    },
    "PaymentFailed": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:handlePaymentFailure",
      "Next": "OrderFailed"
    },
    "OrderFailed": {
      "Type": "Fail",
      "Error": "OrderFailed",
      "Cause": "Order processing failed"
    }
  }
}
json
{
  "Comment": "Order processing workflow",
  "StartAt": "ValidateOrder",
  "States": {
    "ValidateOrder": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:validateOrder",
      "Next": "CheckInventory",
      "Catch": [
        {
          "ErrorEquals": ["InvalidOrder"],
          "Next": "OrderFailed"
        }
      ]
    },
    "CheckInventory": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:checkInventory",
      "Next": "InventoryDecision"
    },
    "InventoryDecision": {
      "Type": "Choice",
      "Choices": [
        {
          "Variable": "$.inStock",
          "BooleanEquals": true,
          "Next": "ProcessPayment"
        }
      ],
      "Default": "OutOfStock"
    },
    "ProcessPayment": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:processPayment",
      "Next": "PaymentDecision",
      "Retry": [
        {
          "ErrorEquals": ["PaymentError"],
          "IntervalSeconds": 2,
          "MaxAttempts": 3,
          "BackoffRate": 2.0
        }
      ]
    },
    "PaymentDecision": {
      "Type": "Choice",
      "Choices": [
        {
          "Variable": "$.paymentApproved",
          "BooleanEquals": true,
          "Next": "ShipOrder"
        }
      ],
      "Default": "PaymentFailed"
    },
    "ShipOrder": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:shipOrder",
      "Next": "NotifyCustomer"
    },
    "NotifyCustomer": {
      "Type": "Task",
      "Resource": "arn:aws:states:::sns:publish",
      "Parameters": {
        "TopicArn": "arn:aws:sns:region:account:order-updates",
        "Message": {
          "orderId.$": "$.orderId",
          "status": "shipped"
        }
      },
      "Next": "OrderSuccess"
    },
    "OrderSuccess": {
      "Type": "Succeed"
    },
    "OutOfStock": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:notifyOutOfStock",
      "Next": "OrderFailed"
    },
    "PaymentFailed": {
      "Type": "Task",
      "Resource": "arn:aws:lambda:region:account:function:handlePaymentFailure",
      "Next": "OrderFailed"
    },
    "OrderFailed": {
      "Type": "Fail",
      "Error": "OrderFailed",
      "Cause": "Order processing failed"
    }
  }
}

4. Monitoring and Observability

4. 监控与可观测性

python
undefined
python
undefined

Monitoring helper

Monitoring helper

import json import logging from aws_lambda_powertools import Logger, Tracer, Metrics from aws_lambda_powertools.utilities.typing import LambdaContext
logger = Logger() tracer = Tracer() metrics = Metrics()
@logger.inject_lambda_context @tracer.capture_lambda_handler def handler(event: dict, context: LambdaContext) -> dict: try: logger.info("Processing event", extra={"event": event})
    # Add custom metrics
    metrics.add_metric(
        name="OrderProcessed",
        unit="Count",
        value=1
    )
    metrics.add_metric(
        name="OrderAmount",
        unit="None",
        value=event.get('amount', 0)
    )

    # Business logic
    result = process_order(event)

    logger.info("Order processed successfully", extra={"orderId": result['orderId']})
    return result

except Exception as e:
    logger.exception("Error processing order")
    metrics.add_metric(
        name="OrderFailed",
        unit="Count",
        value=1
    )
    raise

finally:
    metrics.flush()
def process_order(event): return {"orderId": event.get("id"), "status": "completed"}
undefined
import json import logging from aws_lambda_powertools import Logger, Tracer, Metrics from aws_lambda_powertools.utilities.typing import LambdaContext
logger = Logger() tracer = Tracer() metrics = Metrics()
@logger.inject_lambda_context @tracer.capture_lambda_handler def handler(event: dict, context: LambdaContext) -> dict: try: logger.info("Processing event", extra={"event": event})
    # Add custom metrics
    metrics.add_metric(
        name="OrderProcessed",
        unit="Count",
        value=1
    )
    metrics.add_metric(
        name="OrderAmount",
        unit="None",
        value=event.get('amount', 0)
    )

    # Business logic
    result = process_order(event)

    logger.info("Order processed successfully", extra={"orderId": result['orderId']})
    return result

except Exception as e:
    logger.exception("Error processing order")
    metrics.add_metric(
        name="OrderFailed",
        unit="Count",
        value=1
    )
    raise

finally:
    metrics.flush()
def process_order(event): return {"orderId": event.get("id"), "status": "completed"}
undefined

Best Practices

最佳实践

✅ DO

✅ 建议

  • Design idempotent functions
  • Use event sources efficiently
  • Implement proper error handling
  • Monitor with CloudWatch/Application Insights
  • Use infrastructure as code
  • Implement distributed tracing
  • Version functions for safe deployments
  • Use environment variables for configuration
  • 设计幂等函数
  • 高效使用事件源
  • 实现完善的错误处理
  • 使用CloudWatch/Application Insights进行监控
  • 使用基础设施即代码
  • 实现分布式追踪
  • 为函数添加版本以保障安全部署
  • 使用环境变量存储配置

❌ DON'T

❌ 不建议

  • Create long-running functions
  • Store state in functions
  • Ignore cold start optimization
  • Use synchronous chains
  • Skip testing
  • Hardcode configuration
  • Deploy without monitoring
  • 创建长时间运行的函数
  • 在函数中存储状态
  • 忽略冷启动优化
  • 使用同步链式调用
  • 跳过测试环节
  • 硬编码配置信息
  • 未配置监控就部署

Architecture Patterns

架构模式

  • Event sourcing for audit trails
  • CQRS for read-write optimization
  • Saga pattern for distributed transactions
  • Dead letter queues for failure handling
  • Fan-out/fan-in for parallel processing
  • Circuit breaker for resilience
  • 用于审计追踪的事件溯源
  • 用于读写优化的CQRS
  • 用于分布式事务的Saga模式
  • 用于故障处理的死信队列
  • 用于并行处理的扇出/扇入
  • 用于弹性恢复的断路器

Resources

资源