什么是 Redis-MQ

Redis-MQ 是利用redis实现mq的功能的中间件

特点

无侵入延时队列：无需任何配置和额外的插件即可使用延时队列
兼容spring事务和seata事务：支持消息在事务提交后发送
优雅的api封装: 对外暴露的api及其简单易用
支持消息的负载均衡：通过redis定时轮询注册心跳
支持顺序消息：只需要把消费者和虚拟队列都设置为1
支持定时消息：可以让消息在指定时间执行
支持生产者消费者回调：消息发送结果通过回调通知用户

引入

      <dependency>
            <groupId>io.github.zhaohaoh</groupId>
            <artifactId>redismq-spring-boot-autoconfigure</artifactId>
            <version>Latest Version</version>
        </dependency>

快速开始

application.peoperties配置

#指定环境隔离的分组
spring.redismq.group=samples
spring.redismq.client.host=localhost
#默认的database
spring.redismq.client.database=6
spring.redismq.client.port=6379
spring.redismq.client.username=你的用户
spring.redismq.client.password=你的密码
#虚拟队列数量默认是1，单机redis配多了没有意义
spring.redismq.virtual=1
#默认值true 事务提交后发送
spring.redismq.global-config.send-after-commit=true
#如果有seata事务需要开启 默认值false
spring.redismq.global-config.seata-state=true

案例代码

/**
 * @Author: hzh
 * @Date: 2022/12/26 17:54
 * 生产消息的例子
 */
@RestController
@RequestMapping("producer")
public class ProducerController {
    @Autowired
    private RedisMQTemplate redisMQTemplate;

    /**
     * 发送延迟消息
     */
    @PostMapping("sendDelayMessage")
    public void sendDelayMessage() {
        redisMQTemplate.sendDelayMessage("延时消息消费", "delaytest1", Duration.ofSeconds(60));
    }


    /**
     * 发送普通消息
     */
    @PostMapping("sendMessage")
    public void sendMessage() {
        redisMQTemplate.sendMessage("普通消息消费", "test1");
    }

    /**
     * 发送顺序消息
     */
    @PostMapping("sendOrderMessage")
    public void sendOrderMessage() {
        redisMQTemplate.sendMessage("顺序消息消费", "order");
    }

    /**
     * 发送定时消费消息 带tag
     */
    @PostMapping("sendTimingMessage")
    public void sendTimingMessage() {
        LocalDateTime time = LocalDateTime.of(2022, 12, 26, 14, 20, 30);
        long l = time.toInstant(ZoneOffset.ofHours(8)).toEpochMilli();
        redisMQTemplate.sendTimingMessage("定时消息消费", "time", "bussiness1", l);
    }


    /**
     * 发送定时消费消息 带tag
     */
    @PostMapping("sendMultiTagMessage")
    public void sendMultiTagMessage() {
            redisMQTemplate.sendMessage("多个标签同一topic消息消费1", "MultiTag", "bussiness1");
            redisMQTemplate.sendMessage("多个标签同一topic消息消费2", "MultiTag", "bussiness2");
            redisMQTemplate.sendMessage("多个标签同一topic消息消费1", "MultiTag", "bussiness1");
            redisMQTemplate.sendMessage("多个标签同一topic消息消费2", "MultiTag", "bussiness2");
            redisMQTemplate.sendMessage("多个标签同一topic消息消费1", "MultiTag", "bussiness1");
            redisMQTemplate.sendMessage("多个标签同一topic消息消费2", "MultiTag", "bussiness2");
            redisMQTemplate.sendMessage("多个标签同一topic消息消费1", "MultiTag", "bussiness1");
            redisMQTemplate.sendMessage("多个标签同一topic消息消费2", "MultiTag", "bussiness2");
    }
}

/**
 * @Author: hzh
 * @Date: 2022/12/26 17:54
 * 消费者简单案例
 */
@Component
public class SamplesConsumer {


    /**
     * delaytest1消费延时队列
     */
    @RedisListener(topic = "delaytest1", delay = true)
    public void delaytest1(String test) {
        System.out.println(test);
    }

    /**
     * 普通消息消费
     */
    @RedisListener(topic = "test1")
    public void test1(String test) {
        System.out.println(test);
    }

    /**
     * 顺序消息消费  虚拟队列，消费者线程都设置为1即可保证顺序
     */
    @RedisListener(topic = "order", virtual = 1, concurrency = 1, maxConcurrency = 1)
    public void order(Message message) {
        System.out.println(message);
    }

    @RedisListener(topic = "time",tag = "bussiness1",delay = true)
    public void time(Message message) {
        System.out.println(message);
    }


    /**
     * 多标签同topic消费，会由同一个线程池进行消费
     *
     * @param message 消息
     */
    @RedisListener(topic = "MultiTag",tag = "bussiness1")
    public void multiTag1(Message message) {
        //模拟业务消费
        try {
            Thread.sleep(1000L);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
        String name = Thread.currentThread().getName();
        System.out.println(name+message);
    }

    @RedisListener(topic = "MultiTag",tag = "bussiness2")
    public void multiTag2(Message message) {
        //模拟业务消费
        try {
            Thread.sleep(1000L);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
        String name = Thread.currentThread().getName();
        System.out.println(name+message);
    }
}

如何处理消息堆积

由于内存队列的特性，无法堆积消息(消息累积到一定数量会拒绝)。因此框架提供了发送消息的生产者前后回调和消费者前后回调。默认消费失败加入redis的死信队列默认生产发消息失败打印失败日志可自定义实现对消息持久化mysql等第三方存储库

@Configuration
public class RedisMQInterceptorConfiguration {
    
    @Bean 
    public ConsumeInterceptor redisDeadQueueHandleInterceptor() {
        return new 自定义ConsumeInterceptor实现类();
    }
    
    @Bean 
    public ProducerInterceptor producerInterceptor() {
        return new 自定义ProducerInterceptor实现类();
    }
}

作者

微信:huangzhaohao1995

版权 | License

Apache License 2.0

dsapr/redis-mq

什么是 Redis-MQ

特点

引入

快速开始

application.peoperties配置

案例代码

如何处理消息堆积

作者

版权 | License