Kafka接收消息

// 采用监听得方式接收 @Payload标记消息体内容.
@KafkaListener(topics = {"test"},groupId = "hello")
public void onEvent(@Payload String event,
                   @Header(value = KafkaHeaders.RECEIVED_TOPIC) String topic,
                   @Header(value = KafkaHeaders.RECEIVED_PARTITION_ID) String partition){
   System.out.println("读取到了时间消息: " + event);
}

Acknowledgment

开启手动确认模式；

listener:
	ack-mode: manual

// 采用监听得方式接收 @Payload标记消息体内容.
@KafkaListener(topics = {"test"},groupId = "hello")
public void onEvent(@Payload String event,
                   @Header(value = KafkaHeaders.RECEIVED_TOPIC) String topic,
                   @Header(value = KafkaHeaders.RECEIVED_PARTITION_ID) String partition,
                   ConsumerRecord<String,String> record,
                   Acknowledgment ack){
   ack.acknowledge(); // 手动确认，告诉kafka服务器该消息我已经收到了. 
   System.out.println("读取到了时间消息: " + event);
}

读消息指定分区

@KafkaListener(groupId = "hello",
           topicPartitions = {
               @TopicPartition(
                       topic = "${kafka.topic.test}",
                       partitions = {"0","1","2"}, // 0 1 2分区不限制偏移量
                       partitionOffsets = { // 3 分区只读 3偏移量之后的; 4分区只读 4偏移量之后的
                               @PartitionOffset(partition = "3",initialOffset = "3"),
                               @PartitionOffset(partition = "4",initialOffset = "3")
                       })
           }
)

批量消费

修改配置

kafka:
    bootstrap-servers: 192.168.225.128:9092
    listener:
      type: batch
    # 每次读取20条
    consumer:
      max-poll-records: 20

消费者端接收一个List即可

@KafkaListener(topics = {"hi"},groupId = "batchGroup2")
public void onEvent3(List<ConsumerRecord<String,String>> records){
    System.out.println(records.size());
}

消息拦截