Spring Cloud Stream Kafka

Spring Cloud Stream Kafka 基本使用

RocketMQ对Spring Cloud Stream 的介绍

Spring Cloud Stream 体系及原理介绍

主要概念

应⽤用模型
Binder 抽象
持久化发布/订阅⽀支持
消费分组⽀支持
分区⽀支持

基本概念

Source：Stream 发送源

Sink：Stream 接收器器

Processor:

生产者

配置

spring:
  application:
    name: stream-sink
  cloud:
    stream:
      kafka:
        binder:
          brokers: localhost:9092
      bindings:
        goods-out: # 输出通道
          destination: goods  # 对应的topic
          contentType: application/json
        #也可以bing多个通道
        log-out:
            destination: log  # 对应的topic
          contentType: application/json
      default-binder: kafka  #与consul 使用时需要指定 binder

定义通道

public interface GreetingsStreams {
    String OUTPUT = "goods-out"; // 与配置中一样

    @Output(OUTPUT)
    MessageChannel outboundGreetings();
}

激活

@EnableBinding(GreetingsStreams.class)
public class StreamsConfig {
}

发送消息

@Service
@Slf4j
public class GreetingsService {
    private final GreetingsStreams greetingsStreams;

    public GreetingsService(GreetingsStreams greetingsStreams) {
        this.greetingsStreams = greetingsStreams;
    }

    public void sendGreeting(final Greetings greetings) {
        log.info("Sending greetings {}", greetings);

        MessageChannel messageChannel = greetingsStreams.outboundGreetings();
        messageChannel.send(MessageBuilder
                .withPayload(greetings)
                .setHeader(MessageHeaders.CONTENT_TYPE, MimeTypeUtils.APPLICATION_JSON)
                .build());
    }
}

消费者

配置

spring:
  application:
    name: stream-sink

  cloud:
    stream:
      kafka:
        binder:
          brokers: localhost:9092
      bindings:
        goods-in:
          destination: goods
          contentType: application/json
          group: finance # 指定消费者组
      default-binder: kafka

定义通道

public interface GreetingsStreams {
    String INPUT = "goods-in";
    @Input(INPUT)
    SubscribableChannel inboundGreetings();
}

激活

@EnableBinding(GreetingsStreams.class)
public class StreamsConfig {
}

接收消息

@Component
@Slf4j
public class GreetingsListener {
    /**
     *
     * @param greetings
     * @param partition  从哪个分区获取的数据
     */
    @StreamListener(GreetingsStreams.INPUT)
    public void handleGreetings(@Payload Greetings greetings,@Header(KafkaHeaders.RECEIVED_PARTITION_ID) int partition) {
        log.info("Received message: {},from partition : {}", greetings,partition);
    }
}

消费分区

kafka 的partition 是一个有序队列，指定key可以将相同key的数据发送到同一个partition，可以保证消息有序消费

spring:
  application:
    name: stream-source

  cloud:
    stream:
      kafka:
        binder:
          brokers: localhost:9092
          auto-add-partitions: true
      bindings:
        goods-out:
          destination: goods
          contentType: application/json
          producer:
            partition-key-expression:  headers['partitionKey'] # partition key 表达式
            partition-count: 4  # partition 数量
      default-binder: kafka

发送端

private final static String PARTITION_KEY = "partitionKey";
private final GreetingsStreams greetingsStreams;

public GreetingsService(GreetingsStreams greetingsStreams) {
   this.greetingsStreams = greetingsStreams;
}

public void sendGreeting(final Greetings greetings, String key) {
    log.info("Sending greetings {}", greetings);

    MessageChannel messageChannel = greetingsStreams.outboundGreetings();
    messageChannel.send(MessageBuilder
            .withPayload(greetings)
            .setHeader(MessageHeaders.CONTENT_TYPE, MimeTypeUtils.APPLICATION_JSON)
            .setHeader(PARTITION_KEY, key)
            .build());
}

接收端

@StreamListener( target = GreetingsStreams.INPUT, condition = "headers['partitionKey']=='2'")
可以指定condition，接收指定条件的消息

    @StreamListener( target = GreetingsStreams.INPUT, condition = "headers['partitionKey']=='1'")

    public void handleKey1(@Payload Greetings greetings,@Header(KafkaHeaders.RECEIVED_PARTITION_ID) int partition) {
        log.info("Received message: {},from partition : {}", greetings,partition);
    }

    @StreamListener( target = GreetingsStreams.INPUT, condition = "headers['partitionKey']=='2'")

    public void handleKey2(@Payload Greetings greetings,@Header(KafkaHeaders.RECEIVED_PARTITION_ID) int partition) {
        log.info("Received message: {},from partition : {}", greetings,partition);
    }


    @StreamListener( target = GreetingsStreams.INPUT)

    public void handle(@Payload Greetings greetings,@Header(KafkaHeaders.RECEIVED_PARTITION_ID) int partition) {
        log.info("Received message: {},from partition : {}", greetings,partition);
    }

手动应答

消费者端：

可以设置 spring.cloud.stream.kafka.bindings.input.consumer.autoCommitOffset 为false

@SpringBootApplication
@EnableBinding(Sink.class)
public class ManuallyAcknowdledgingConsumer {

 public static void main(String[] args) {
     SpringApplication.run(ManuallyAcknowdledgingConsumer.class, args);
 }

 @StreamListener(Sink.INPUT)
 public void process(Message message) {
     Acknowledgment acknowledgment = message.getHeaders().get(KafkaHeaders.ACKNOWLEDGMENT, Acknowledgment.class);
     if (acknowledgment != null) {
         System.out.println("Acknowledgment provided");
         acknowledgment.acknowledge();
     }
 }
}

测试不成功，控制台输出的consumerConfig 依然为true,不知道为啥

auto.commit.interval.ms = 5000
    auto.offset.reset = latest
    bootstrap.servers = [localhost:9092]
    check.crcs = true
    client.id = 
    connections.max.idle.ms = 540000
    default.api.timeout.ms = 60000
    enable.auto.commit = true  一直为自动提交

生产端可以将应答配置为同步的

spring.cloud.stream.kafka.bindings.output.producer.sync=true

https://cloud.spring.io/spring-cloud-static/spring-cloud-stream-binder-kafka/3.0.0.RELEASE/reference/html/spring-cloud-stream-binder-kafka.html

https://docs.spring.io/spring-cloud-stream/docs/current/reference/htmlsingle/

Spring Cloud Stream Kafka