Spring Cloud Stream簡(jiǎn)單用法

更新時(shí)間：2021年07月23日 14:50:36 作者：微瞰技術(shù)

Spring cloud stream是為構(gòu)建微服務(wù)消息驅(qū)動(dòng)而產(chǎn)生的一種框架。Spring Cloud Stream基于Spring boot的基礎(chǔ)上，可創(chuàng)建獨(dú)立的、生產(chǎn)級(jí)別的Spring應(yīng)用，并采用Spring Integration來(lái)連接消息中間件提供消息事件驅(qū)動(dòng)，一起看看吧

Spring Cloud Stream對(duì)Spring Cloud體系中的Mq進(jìn)⾏了很好的上層抽象，可以讓我們與具體消息中間件解耦合，屏蔽掉了底層具體MQ消息中間件的細(xì)節(jié)差異，就像Hibernate屏蔽掉了具體數(shù)據(jù)庫(kù)（Mysql/Oracle⼀樣）。如此⼀來(lái)，我們學(xué)習(xí)、開發(fā)、維護(hù)MQ都會(huì)變得輕松。⽬前Spring Cloud Stream原生⽀持RabbitMQ和Kafka，阿里在這個(gè)基礎(chǔ)上提供了RocketMQ的支持

簡(jiǎn)單使用Spring Cloud Stream 構(gòu)建基于RocketMQ的生產(chǎn)者和消費(fèi)者

生產(chǎn)者

pom文件中加入依賴

<dependencies>
        <dependency>
            <groupId>org.springframework.cloud</groupId>
            <artifactId>spring-cloud-starter-netflix-eureka-client</artifactId>
        </dependency>
        <dependency>
            <groupId>com.alibaba.cloud</groupId>
            <artifactId>spring-cloud-starter-stream-rocketmq</artifactId>
            <version>2.1.0.RELEASE</version>
        </dependency>

    </dependencies>

配置文件中增加關(guān)于Spring Cloud Stream binder和bindings的配置

spring:
  application:
    name: zhao-cloud-stream-producer
  cloud:
    stream:
      rocketmq:
        binder:
          name-server: 127.0.0.1:9876
        bindings:
          output:
            producer:
              group: test
              sync: true
      bindings:
        output:
          destination: stream-test-topic
          content-type: text/plain # 內(nèi)容格式。這里使用 JSON

其中destination代表生產(chǎn)的數(shù)據(jù)發(fā)送到的topic 然后定義一個(gè)channel用于數(shù)據(jù)發(fā)送

import org.springframework.cloud.stream.annotation.Output;
import org.springframework.messaging.MessageChannel;

public interface TestChannel {
    @Output("output")
    MessageChannel output();
}

最后構(gòu)造數(shù)據(jù)發(fā)送的接口

@Controller
public class SendMessageController {
    @Resource
    private TestChannel testChannel;

    @ResponseBody
    @RequestMapping(value = "send", method = RequestMethod.GET)
    public String sendMessage() {
        String messageId = UUID.randomUUID().toString();
        Message<String> message = MessageBuilder
                .withPayload("this is a test:" + messageId)
                .setHeader(MessageConst.PROPERTY_TAGS, "test")
                .build();
        try {
            testChannel.output().send(message);
            return messageId + "發(fā)送成功";
        } catch (Exception e) {
            return messageId + "發(fā)送失敗，原因：" + e.getMessage();
        }
    }
}

消費(fèi)者

消費(fèi)者的pom引入與生產(chǎn)者相同，在此不再贅述，配置時(shí)需要將stream的output修改為input并修改對(duì)應(yīng)屬性

spring:
  application:
    name: zhao-cloud-stream-consumer
  cloud:
    stream:
      rocketmq:
        binder:
          name-server: 127.0.0.1:9876
        bindings:
          input:
            consumer:
              tags: test
      bindings:
        input:
          destination: stream-test-topic
          content-type: text/plain # 內(nèi)容格式。這里使用 JSON
          group: test

另外關(guān)于channel的構(gòu)造也要做同樣的修改

import org.springframework.cloud.stream.annotation.Input;
import org.springframework.messaging.SubscribableChannel;

public interface TestChannel {
    @Input("input")
    SubscribableChannel input();
}

最后我在啟動(dòng)類中對(duì)收到的消息進(jìn)行了監(jiān)聽

   @StreamListener("input")
    public void receiveInput(@Payload Message message) throws ValidationException {
        System.out.println("input1 receive: " + message.getPayload() + ", foo header: " + message.getHeaders().get("foo"));
    }

測(cè)試結(jié)果

file

Stream其他特性

消息發(fā)送失敗的處理

消息發(fā)送失敗后悔發(fā)送到默認(rèn)的一個(gè)“topic.errors"的channel中(topic是配置的destination)。要配置消息發(fā)送失敗的處理，需要將錯(cuò)誤消息的channel打開消費(fèi)者配置如下

spring:
  application:
    name: zhao-cloud-stream-producer
  cloud:
    stream:
      rocketmq:
        binder:
          name-server: 127.0.0.1:9876
        bindings:
          output:
            producer:
              group: test
              sync: true
      bindings:
        output:
          destination: stream-test-topic
          content-type: text/plain # 內(nèi)容格式。這里使用 JSON
          producer:
            errorChannelEnabled: true

在啟動(dòng)類中配置錯(cuò)誤消息的Channel信息

 @Bean("stream-test-topic.errors")
    MessageChannel testoutPutErrorChannel(){
        return new PublishSubscribeChannel();
    }

新建異常處理service

import org.springframework.integration.annotation.ServiceActivator;
import org.springframework.messaging.Message;
import org.springframework.stereotype.Service;

@Service
public class ErrorProducerService {

    @ServiceActivator(inputChannel = "stream-test-topic.errors")
    public void receiveProducerError(Message message){
        System.out.println("receive error msg :"+message);
    }
}

當(dāng)發(fā)生異常時(shí)，由于測(cè)試類中已經(jīng)將異常捕獲，處理發(fā)送異常主要是在這里進(jìn)行。模擬，應(yīng)用與rocketMq斷開的場(chǎng)景?？梢?/p>

file file

消費(fèi)者錯(cuò)誤處理

首先增加配置為

spring:
  application:
    name: zhao-cloud-stream-producer
  cloud:
    stream:
      rocketmq:
        binder:
          name-server: 127.0.0.1:9876
        bindings:
          output:
            producer:
              group: test
              sync: true
      bindings:
        output:
          destination: stream-test-topic
          content-type: text/plain # 內(nèi)容格式。這里使用 JSON
          producer:
            errorChannelEnabled: true

增加相應(yīng)的模擬異常的操作

 @StreamListener("input")
    public void receiveInput(@Payload Message message) throws ValidationException {
        //System.out.println("input1 receive: " + message.getPayload() + ", foo header: " + message.getHeaders().get("foo"));
        throw new RuntimeException("oops");
    }
    @ServiceActivator(inputChannel = "stream-test-topic.test.errors")
    public void receiveConsumeError(Message message){
        System.out.println("receive error msg"+message.getPayload());
    }

file