Kafka监听器中的钩子

Question

在kafka收听消息之前/之后是否有任何类型的钩子可用？

用例：必须为设置MDC关联id以执行日志跟踪

我在找什么？前/后回调方法，以便可以在进入时设置MDC关联id，并最终在退出时清除MDC。

编辑的场景：我正在获取作为Kafka Headers一部分的相关ID，我想在Kafka Listener中收到消息后立即在MDC中设置相同的ID

感谢您的帮助

Answer 1

你可以在你的侦听器bean中添加一个around建议。。。

@SpringBootApplication
public class So59854374Application {

    public static void main(String[] args) {
        SpringApplication.run(So59854374Application.class, args);
    }

    @Bean
    public static BeanPostProcessor bpp() { // static is important
        return new BeanPostProcessor() {

            @Override
            public Object postProcessBeforeInitialization(Object bean, String beanName) throws BeansException {
                if (bean instanceof MyListener) {
                    ProxyFactoryBean pfb = new ProxyFactoryBean();
                    pfb.setTarget(bean);
                    pfb.addAdvice(new MethodInterceptor() {

                        @Override
                        public Object invoke(MethodInvocation invocation) throws Throwable {
                            try {
                                System.out.println("Before");
                                return invocation.proceed();
                            }
                            finally {
                                System.out.println("After");
                            }
                        }

                    });
                    return pfb.getObject();
                }
                return bean;
            }

        };
    }

    @Bean
    public NewTopic topic() {
        return TopicBuilder.name("so59854374").partitions(1).replicas(1).build();
    }

    @Bean
    public ApplicationRunner runner(KafkaTemplate<String, String> template) {
        return args -> template.send("so59854374", "foo");
    }

}

@Component
class MyListener {

    @KafkaListener(id = "so59854374", topics = "so59854374")
    public void listen(String in) {
        System.out.println(in);
    }

}

和

Before
foo
After

编辑

如果将字节[]mdc作为附加参数添加到kafka侦听器方法中，则可以在调用中使用getArguments（）[1]。

另一种解决方案是向侦听器容器工厂添加一个RecordInterceptor，它允许您在原始消费记录传递到侦听器适配器之前访问它。

/**
 * An interceptor for {@link ConsumerRecord} invoked by the listener
 * container before invoking the listener.
 *
 * @param <K> the key type.
 * @param <V> the value type.
 *
 * @author Gary Russell
 * @since 2.2.7
 *
 */
@FunctionalInterface
public interface RecordInterceptor<K, V> {

    /**
     * Perform some action on the record or return a different one.
     * If null is returned the record will be skipped.
     * @param record the record.
     * @return the record or null.
     */
    @Nullable
    ConsumerRecord<K, V> intercept(ConsumerRecord<K, V> record);

}

/**
 * Set an interceptor to be called before calling the listener.
 * Does not apply to batch listeners.
 * @param recordInterceptor the interceptor.
 * @since 2.2.7
 */
public void setRecordInterceptor(RecordInterceptor<K, V> recordInterceptor) {
    this.recordInterceptor = recordInterceptor;
}

如果您使用的是批处理侦听器，Kafka提供了一个消费者接收器。