我已经将Spring Kafka监听器初始化为

@Bean
public Map<String, Object> consumerConfig() {
    final HashMap<String, Object> result = new HashMap<>();
    result.put(BOOTSTRAP_SERVERS_CONFIG, bootstrapServers);
    result.put(GROUP_ID_CONFIG, groupId);
    result.put(KEY_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class);
    result.put(VALUE_DESERIALIZER_CLASS_CONFIG, MyKafkaJacksonRulesExecutionResultsDeserializer.class);
    return result;
}

@Bean
public ConsumerFactory<Long, MessageResult> consumerFactory() {
    return new DefaultKafkaConsumerFactory<>(consumerConfig());
}

@Bean
public ConcurrentKafkaListenerContainerFactory<Long, MessageResult> kafkaListenerContainerFactory() {
    ConcurrentKafkaListenerContainerFactory<Long, MessageResult> containerFactory = new ConcurrentKafkaListenerContainerFactory<>();
    containerFactory.setConsumerFactory(consumerFactory());
    containerFactory.setConcurrency(KAFKA_LISTENER_THREADS_COUNT);
    containerFactory.getContainerProperties().setPollTimeout(KAFKA_LISTENER_POLL_TIMEOUT);
    containerFactory.getContainerProperties().setAckOnError(true);
    containerFactory.getContainerProperties().setAckMode(RECORD);
    return containerFactory;
}

并用作
@KafkaListener(topics = "${spring.kafka.out-topic}")
public void processSrpResults(MessageResult result) {

反序列化器在反序列化过程中引发异常,这会导致无限循环,因为监听器无法获取消息。

我怎样才能使kafka监听器在错误时提交?

最佳答案

我创建了反序列化器的子类,该子类引发了异常。然后,在我的配置中将其用作解串器。然后,您的处理器必须处理空对象。

public class MyErrorHandlingDeserializer extends ExceptionThrowingDeserializer {

    @Override
    public Object deserialize(String topic, byte[] data) {
        try {
            return super.deserialize(topic, data);
        } catch (Exception e) {
            log.error("Problem deserializing data " + new String(data) + " on topic " + topic, e);
            return null;
        }
    }
}

关于Spring Kafka监听器无限循环出错,我们在Stack Overflow上找到一个类似的问题:https://stackoverflow.com/questions/44796543/

10-13 04:38