直接上代码

pom文件

<dependency>
    <groupId>org.springframework.kafka</groupId>
    <artifactId>spring-kafka</artifactId>
    <version>2.6.10</version>
</dependency>

application.yml设置auto-offset-reset: earliest，如果不设置这个值，默认的是latest

spring:
  application:
    name: sd-km

  kafka:
    bootstrap-servers: 192.188.234.12:9092
    consumer:
      auto-offset-reset: earliest
      group-id: test

auto-offset-reset值的含义解释

• earliest
• 当各分区下有已提交的offset时，从提交的offset开始消费；无提交的offset时，从头开始消费
• latest
• 当各分区下有已提交的offset时，从提交的offset开始消费；无提交的offset时，消费新产生的该分区下的数据
• none
• topic各分区都存在已提交的offset时，从offset后开始消费；只要有一个分区不存在已提交的offset，则抛出异常
   

工具类：

@Component
public class KafkaOffsetTools {

    private static KafkaOffsetTools _this;

    @Autowired
    private ConsumerFactory<Long, String> consumerFactory;

    @Autowired
    private KafkaTemplate<String, String> kafkaTemplate;

    @PostConstruct
    public void init() {
        _this = this;
    }

    /**
     * 获取lag
     **/
    public static Long getLAG(String topic) {
        return getDescribe(topic)[2];
    }

    /**
     * 获取分组下的表述信息
     **/
    private static long[] getDescribe(String topic) {
        long[] describe = new long[3];
        Consumer<Long, String> consumer = createConsumer();

        List<PartitionInfo> partitionInfos = _this.kafkaTemplate.partitionsFor(topic);
        List<TopicPartition> tp = new ArrayList<>();
        partitionInfos.forEach(str -> {
            TopicPartition topicPartition = new TopicPartition(topic, str.partition());
            tp.add(topicPartition);
            long logEndOffset = consumer.endOffsets(tp).get(topicPartition);

            consumer.assign(tp);
            //consumer.position(topicPartition);
            long currentOffset = consumer.position(topicPartition);

            System.out.println("logEndOffset : " + logEndOffset + ", currentOffset : "+ currentOffset);
            describe[0] += currentOffset;
            describe[1] += logEndOffset;
            describe[2] = describe[1] - describe[0];

            tp.clear();
        });

        System.out.println(Arrays.toString(describe));
        return describe;
    }

    /**
     * 创建消费者
     **/
    private static Consumer<Long, String> createConsumer() {
        return _this.consumerFactory.createConsumer();
    }
}

在需要的地方调用：

KafkaOffsetTools.getLAG("topicname")