消费kafka数据：处理消息模块改成为多线程的实现方式

最新推荐文章于 2026-05-22 10:13:00 发布

原创最新推荐文章于 2026-05-22 10:13:00 发布 · 454 阅读

0 ·

本内容遵循CC 4.0 BY-SA版权协议

标签

#kafka

kafka 专栏收录该内容

3 篇文章

订阅专栏

本文介绍了一种使用多线程处理Kafka消息的消费者实现方式，通过自定义KafkaConsumerThread类和RecordsHandler类，实现了消息的并行处理和消费位移的同步更新。代码示例展示了如何配置消费者属性，创建线程池，以及处理消费记录。

package application;

import org.apache.kafka.clients.consumer.*;
import org.apache.kafka.common.TopicPartition;
import org.apache.kafka.common.serialization.StringDeserializer;

import java.time.Duration;
import java.util.Collections;
import java.util.List;
import java.util.Map;
import java.util.Properties;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.ThreadPoolExecutor;
import java.util.concurrent.TimeUnit;

/**
 *多线程消费实现方式，就是在处理消息模块改成多线程的实现方式
 * */
public class ThirdMultiConsumerThreadDemo {

    public static final String brokers = "localhost:9092";
    public static final String topic = "topic-spark";
    public static final String groupId = "group-spark";

    //每一个处理消息的RecordsHandler类在处理完消息之后都将对应的消费位移保存到共享变量offsets中
    public static Map<TopicPartition, OffsetAndMetadata> offsets;

    public static Properties initConfig(){
        Properties props = new Properties();

        props.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class.getName());
        props.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG,StringDeserializer.class.getName());
        props.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG,brokers);
        props.put(ConsumerConfig.GROUP_ID_CONFIG,groupId);
        props.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG,true);
        return props;
    }

    public static void main(String[] args) {
        Properties props = initConfig();
        KafkaConsumerThread comsumerThread = new KafkaConsumerThread(
          props,
          topic,
          Runtime.getRuntime().availableProcessors());

        comsumerThread.start();
    }

    public static class KafkaConsumerThread extends Thread{
        private KafkaConsumer<String,String> kafkaConsumer;
        private ExecutorService executorService;
        private int threadNumber;



        public KafkaConsumerThread(Properties props, String topic, int threadNumber){

            kafkaConsumer = new KafkaConsumer<String, String>(props);
            kafkaConsumer.subscribe(Collections.singletonList(topic));
            this.threadNumber = threadNumber;
            executorService = new ThreadPoolExecutor(
                    threadNumber,
                    threadNumber,
                    0L,
                    TimeUnit.MICROSECONDS,
                    new ArrayBlockingQueue<>(100),
                    /**CallerRunsPolicy()：这个参数可以防止线程池的总体消费能力跟不上poll()拉去的能力，从而导致异常现象的发生。*/
                    new ThreadPoolExecutor.CallerRunsPolicy()
            );
        }

        public void run(){
            try{
                while (true){

                    ConsumerRecords<String,String> records = kafkaConsumer.poll(Duration.ofMinutes(100));
                    if(!records.isEmpty()){
                        executorService.submit(new RecordsHandler(records));
                    }

                    //对应的位移提交实现
                    synchronized (offsets){
                        if(!offsets.isEmpty()){
                            kafkaConsumer.commitSync(offsets);
                            offsets.clear();
                        }
                    }
                }
            }catch (Exception e){
                e.printStackTrace();
            }finally {
                kafkaConsumer.close();
            }
        }
    }

    public static class RecordsHandler extends Thread{
        public final ConsumerRecords<String,String> records ;

        public RecordsHandler(ConsumerRecords<String,String> records){
            this.records = records;
        }

        public void run(){

            for(TopicPartition tp: records.partitions()){
                List<ConsumerRecord<String,String>> tpRecords = records.records(tp);
                //业务处理：tpRecords

                
                long lastConsumerOffset = tpRecords.get(tpRecords.size() -1).offset();
                //对offsets加锁，防止出现并发问题
                synchronized (offsets){
                    if(!offsets.containsKey(tp)){
                        offsets.put(tp,new OffsetAndMetadata(lastConsumerOffset +1));
                    }else {
                        long position = offsets.get(tp).offset();
                        if(position<lastConsumerOffset+1){
                            offsets.put(tp,new OffsetAndMetadata(lastConsumerOffset+1));
                        }
                    }
                }

            }
        }
    }

}

该代码未解决异常情况下的位移覆盖问题。