消息队列设计：从同步到异步的性能突破-育师

前言

2024年初，我们的订单系统经常出现"超时"问题。用户下单后，系统需要同时调用库存服务、支付服务、通知服务，任何一个服务慢都会导致整个请求超时。

我们决定引入消息队列，将同步调用改为异步处理。这个改造带来了显著的性能提升。

一、问题：同步调用的瓶颈

原始的订单流程是这样的：

python

@app.route('/api/orders', methods=['POST']) def create_order(): # 1. 创建订单 order = Order.create(request.json) # 2. 同步调用库存服务 inventory_response = requests.post( 'http://inventory-service/deduct', json={'product_id': order.product_id, 'quantity': order.quantity} ) if inventory_response.status_code != 200: return {"error": "库存不足"}, 400 # 3. 同步调用支付服务 payment_response = requests.post( 'http://payment-service/pay', json={'order_id': order.id, 'amount': order.amount} ) if payment_response.status_code != 200: return {"error": "支付失败"}, 400 # 4. 同步调用通知服务 notify_response = requests.post( 'http://notify-service/send', json={'order_id': order.id, 'type': 'order_created'} ) return {"order_id": order.id}, 201

问题：

任何一个服务慢都会导致整个请求慢；
任何一个服务故障都会导致订单创建失败；
耦合度太高，难以扩展。

性能数据：

库存服务：200ms
支付服务：300ms
通知服务：150ms
总耗时：200 + 300 + 150 =650ms

二、解决方案：引入RabbitMQ

我们选择RabbitMQ作为消息队列。改造后的流程：

2.1 发布订单创建事件

python

import pika import json def create_order(): # 1. 创建订单 order = Order.create(request.json) # 2. 发布事件到消息队列 connection = pika.BlockingConnection(pika.ConnectionParameters('rabbitmq')) channel = connection.channel() # 声明交换机和队列 channel.exchange_declare(exchange='orders', exchange_type='topic') # 发布消息 message = { 'order_id': order.id, 'product_id': order.product_id, 'quantity': order.quantity, 'amount': order.amount } channel.basic_publish( exchange='orders', routing_key='order.created', body=json.dumps(message) ) connection.close() # 立即返回响应 return {"order_id": order.id}, 201

耗时：仅需10ms（发布到队列）

2.2 消费者：库存服务

python

def inventory_consumer(): connection = pika.BlockingConnection(pika.ConnectionParameters('rabbitmq')) channel = connection.channel() channel.exchange_declare(exchange='orders', exchange_type='topic') result = channel.queue_declare(queue='inventory_queue', durable=True) queue_name = result.method.queue # 绑定队列到交换机 channel.queue_bind( exchange='orders', queue=queue_name, routing_key='order.created' ) def callback(ch, method, properties, body): message = json.loads(body) try: # 扣减库存 deduct_inventory( message['product_id'], message['quantity'] ) # 确认消息 ch.basic_ack(delivery_tag=method.delivery_tag) except Exception as e: # 拒绝消息，重新入队 ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True) channel.basic_consume( queue=queue_name, on_message_callback=callback ) print('库存服务已启动，等待消息...') channel.start_consuming() if __name__ == '__main__': inventory_consumer()

2.3 消费者：支付服务

python

def payment_consumer(): connection = pika.BlockingConnection(pika.ConnectionParameters('rabbitmq')) channel = connection.channel() channel.exchange_declare(exchange='orders', exchange_type='topic') result = channel.queue_declare(queue='payment_queue', durable=True) queue_name = result.method.queue channel.queue_bind( exchange='orders', queue=queue_name, routing_key='order.created' ) def callback(ch, method, properties, body): message = json.loads(body) try: # 处理支付 process_payment( message['order_id'], message['amount'] ) ch.basic_ack(delivery_tag=method.delivery_tag) except Exception as e: ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True) channel.basic_consume( queue=queue_name, on_message_callback=callback ) print('支付服务已启动，等待消息...') channel.start_consuming()

2.4 消费者：通知服务

python

def notify_consumer(): connection = pika.BlockingConnection(pika.ConnectionParameters('rabbitmq')) channel = connection.channel() channel.exchange_declare(exchange='orders', exchange_type='topic') result = channel.queue_declare(queue='notify_queue', durable=True) queue_name = result.method.queue channel.queue_bind( exchange='orders', queue=queue_name, routing_key='order.created' ) def callback(ch, method, properties, body): message = json.loads(body) try: # 发送通知 send_notification( message['order_id'], 'order_created' ) ch.basic_ack(delivery_tag=method.delivery_tag) except Exception as e: ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True) channel.basic_consume( queue=queue_name, on_message_callback=callback ) print('通知服务已启动，等待消息...') channel.start_consuming()

三、可靠性保证

3.1 消息持久化

python

# 声明持久化队列 channel.queue_declare( queue='payment_queue', durable=True # 队列持久化 ) # 发布持久化消息 channel.basic_publish( exchange='orders', routing_key='order.created', body=json.dumps(message), properties=pika.BasicProperties( delivery_mode=2 # 消息持久化 ) )

3.2 消息确认机制

python

Copy code

# 手动确认消息 def callback(ch, method, properties, body): try: process_message(body) ch.basic_ack(delivery_tag=method.delivery_tag) # 确认 except Exception as e: ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True) # 拒绝并重新入队 # 禁用自动确认 channel.basic_consume( queue=queue_name, on_message_callback=callback, auto_ack=False # 手动确认 )

3.3 死信队列

python

# 声明死信交换机 channel.exchange_declare(exchange='dlx', exchange_type='direct') channel.queue_declare(queue='dead_letter_queue', durable=True) channel.queue_bind(exchange='dlx', queue='dead_letter_queue') # 声明普通队列，指定死信交换机 channel.queue_declare( queue='payment_queue', durable=True, arguments={ 'x-dead-letter-exchange': 'dlx', 'x-dead-letter-routing-key': 'dead_letter' } )

四、监控和告警

python

import logging logging.basicConfig(level=logging.INFO) logger = logging.getLogger(__name__) def callback(ch, method, properties, body): start_time = time.time() try: process_message(body) duration = time.time() - start_time logger.info(f"消息处理成功, 耗时: {duration}ms") ch.basic_ack(delivery_tag=method.delivery_tag) except Exception as e: logger.error(f"消息处理失败: {str(e)}") ch.basic_nack(delivery_tag=method.delivery_tag, requeue=True)