springboot:整合rabbitmq之重试机制

x33g5p2x  于2022-06-16 转载在 Spring  
字(5.2k)|赞(0)|评价(0)|浏览(613)

springboot:整合rabbitmq之重试机制

一、项目准备

依赖

<dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-web</artifactId>
        </dependency>

        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-amqp</artifactId>
        </dependency>

配置类

@Configuration
public class RabbitMQConfiguration {

    @Bean
    public ConnectionFactory rabbitConnectionFactory() {
        CachingConnectionFactory connectionFactory = new CachingConnectionFactory();
        String rabbitmqHost = "127.0.0.1";
        String rabbitmqPort = "5672";
        String rabbitmqUsername = "guest";
        String rabbitmqPassword = "guest";
        String rabbitmqVirtualHost = "/";
        connectionFactory.setHost(rabbitmqHost);
        connectionFactory.setPort(Integer.parseInt(rabbitmqPort));
        connectionFactory.setUsername(rabbitmqUsername);
        connectionFactory.setPassword(rabbitmqPassword);
        connectionFactory.setVirtualHost(rabbitmqVirtualHost);
//        connectionFactory.setPublisherReturns(true);//开启return模式
//        connectionFactory.setPublisherConfirms(true);//开启confirm模式
        return connectionFactory;
    }

    @Bean(name = "rabbitTemplate")
    //必须是prototype类型
    @Scope(ConfigurableBeanFactory.SCOPE_PROTOTYPE)
    public RabbitTemplate rabbitTemplate() {
        return new RabbitTemplate(rabbitConnectionFactory());
    }

    @Bean("customContainerFactory")
    public SimpleRabbitListenerContainerFactory containerFactory(SimpleRabbitListenerContainerFactoryConfigurer configurer, ConnectionFactory connectionFactory) {
        SimpleRabbitListenerContainerFactory factory = new SimpleRabbitListenerContainerFactory();
        //设置线程数
        factory.setConcurrentConsumers(1);
        //最大线程数
        factory.setMaxConcurrentConsumers(1);
//        //设置为手动确认MANUAL(手动),AUTO(自动);
//        factory.setAcknowledgeMode(AcknowledgeMode.MANUAL);
        // 设置prefetch
        factory.setPrefetchCount(1);
        configurer.configure(factory, connectionFactory);
        return factory;
    }
}

二、案例重现

@Slf4j
@RestController
public class RetryController {

    @Autowired
    private RabbitTemplate rabbitTemplate;

    @GetMapping("/retry")
    public void retry() {
        String message = "Hello World !";
        rabbitTemplate.convertAndSend("retry_exchange", "retry_key", message);
        System.out.println(" [ 生产者 ] Sent ==> '" + message + "'");
    }

    private int count = 1;

    @RabbitListener(containerFactory = "customContainerFactory", bindings = {@QueueBinding(
            value = @Queue(value = "retry_a", declare = "true"),
            exchange = @Exchange(value = "retry_exchange", type = ExchangeTypes.DIRECT, durable = "true"),
            key = "retry_key"
    )
    })
    public void retry(Message message) {
        log.info("当前执行次数:{}", count++);
        log.info(" [ 消费者@A号 ] 接收到消息 ==> '" + new String(message.getBody()));
        // 制造异常
        int i = 1 / 0;
        log.info(" [ 消费者@A号 ] 消费了消息 ==> '" + new String(message.getBody()));
    }
}

启动测试:

无限循环报错
停止后,消息重回Ready状态

三、实现消息重试

实现重试

spring:
  rabbitmq:
    listener:
      simple:
        retry:
          enabled: true
          max-attempts: 5 # 重试次数
          max-interval: 10000   # 重试最大间隔时间
          initial-interval: 2000  # 重试初始间隔时间
          multiplier: 2 # 间隔时间乘子,间隔时间*乘子=下一次的间隔时间,最大不能超过设置的最大间隔时间

重启测试

第一次执行时间2s,第二次4s,第三次8s,第四次16s,第五次由于设置了最大间隔为10s,所有变成了10s

最后查看retry_a队列,消息没有了,也就是说重试五次失败之后就会移除该消息

移除操作是由日志中的这个类处理:RejectAndDontRequeueRecoverer(拒绝和不要重新排队)

对重试失败的消息重新排队

使用下 ImmediateRequeueMessageRecoverer 重新排队在RabbitMQConfiguration中配置

@Bean
    public MessageRecoverer messageRecoverer() {
        return new ImmediateRequeueMessageRecoverer();
    }

重启运行:

可以看出:重试5次之后,返回队列,然后再重试5次,周而复始直到不抛出异常为止,这样还是会影响后续的消息消费

把重试失败消息放入重试失败队列

接着使用 RepublishMessageRecoverer 重新发布在RabbitMQConfiguration中配置

public static final String RETRY_FAILURE_KEY = "retry.failure.key";
    public static final String RETRY_EXCHANGE = "retry_exchange";

    //@Bean   这个注释掉了
    public MessageRecoverer messageRecoverer() {
        return new ImmediateRequeueMessageRecoverer();
    }

    @Bean
    public MessageRecoverer messageRecoverer(RabbitTemplate rabbitTemplate) {
        // 需要配置交换机和绑定键
        return new RepublishMessageRecoverer(rabbitTemplate, RETRY_EXCHANGE, RETRY_FAILURE_KEY);
    }

创建重试失败消息监听

@Slf4j
@RestController
public class RetryController {

    @Autowired
    private RabbitTemplate rabbitTemplate;

    @GetMapping("/retry")
    public void retry() {
        String message = "Hello World !";
        rabbitTemplate.convertAndSend("retry_exchange", "retry_key", message);
        System.out.println(" [ 生产者 ] Sent ==> '" + message + "'");
    }

    private int count = 1;

    @RabbitListener(containerFactory = "customContainerFactory", bindings = {@QueueBinding(
            value = @Queue(value = "retry_a", declare = "true"),
            exchange = @Exchange(value = "retry_exchange", type = ExchangeTypes.DIRECT, durable = "true"),
            key = "retry_key"
    )
    })
    public void retry(Message message) {
        log.info("当前执行次数:{}", count++);
        log.info(" [ 消费者@A号 ] 接收到消息 ==> '" + new String(message.getBody()));
        // 制造异常
        int i = 1 / 0;
        log.info(" [ 消费者@A号 ] 消费了消息 ==> '" + new String(message.getBody()));
    }

    @RabbitListener(containerFactory = "customContainerFactory",bindings = @QueueBinding(
            value = @Queue(value = "retry_failure_queue"),
            exchange = @Exchange(value = "retry_exchange"),
            key = "retry.failure.key"
    ))
    public void retryFailure(Message message) {
        log.info(" [ 消费者@重试失败号 ] 接收到消息 ==> '" + new String(message.getBody()));
    }
}

重启,运行结果:

重试5次之后,将消息 Republishing failed message to exchange ‘retry.exchange’ with routing key retry-key 转发到重试失败队列,由重试失败消费者消费

开发者涨薪指南

48位大咖的思考法则、工作方式、逻辑体系

相关文章