spring kafka consumer原理解析一
前言:Spring kafka 是 Spring 对 kafka API的一次封装,省去了写生产和消费的代码,而只需要加个注解即可使用 kafka 收发消息。然而 Spring 是如何实现的呢?与我们自己手写有啥不同?see ↓
以下仅对消费者源码进行分析:
一、消费者的创建
老套路:查看源码首先是找到入口,无可厚非,spring kafka 的入口即为@KafkaListenner注解,因为我们在使用Spring kafka时配置一个@KafkaListnner即可消费到 kafka 的数据。
(1)入口:KafkaListenerAnnotationBeanPostProcessor 一个实现了 BPP 的类,以下只贴出核心的几个方法 ↓
public class KafkaListenerAnnotationBeanPostProcessor<K, V> implements BeanPostProcessor, Ordered, BeanFactoryAware, SmartInitializingSingleton { @Override public Object postProcessAfterInitialization(final Object bean, final String beanName) throws BeansException { if (!this.nonAnnotatedClasses.contains(bean.getClass())) { Class<?> targetClass = AopUtils.getTargetClass(bean); Collection<KafkaListener> classLevelListeners = findListenerAnnotations(targetClass); // 找到所有有@KafkaListenner的方法和类 final boolean hasClassLevelListeners = classLevelListeners.size() > 0; final List<Method> multiMethods = new ArrayList<Method>(); Map<Method, Set<KafkaListener>> annotatedMethods = MethodIntrospector.selectMethods(targetClass, new MethodIntrospector.MetadataLookup<Set<KafkaListener>>() { @Override public Set<KafkaListener> inspect(Method method) { Set<KafkaListener> listenerMethods = findListenerAnnotations(method); return (!listenerMethods.isEmpty() ? listenerMethods : null); } }); if (hasClassLevelListeners) { Set<Method> methodsWithHandler = MethodIntrospector.selectMethods(targetClass, new ReflectionUtils.MethodFilter() { @Override public boolean matches(Method method) { return AnnotationUtils.findAnnotation(method, KafkaHandler.class) != null; } }); multiMethods.addAll(methodsWithHandler); } if (annotatedMethods.isEmpty()) { this.nonAnnotatedClasses.add(bean.getClass()); if (this.logger.isTraceEnabled()) { this.logger.trace("No @KafkaListener annotations found on bean type: " + bean.getClass()); } } else { // Non-empty set of methods for (Map.Entry<Method, Set<KafkaListener>> entry : annotatedMethods.entrySet()) { Method method = entry.getKey(); for (KafkaListener listener : entry.getValue()) { processKafkaListener(listener, method, bean, beanName); } } if (this.logger.isDebugEnabled()) { this.logger.debug(annotatedMethods.size() + " @KafkaListener methods processed on bean '" + beanName + "': " + annotatedMethods); } } if (hasClassLevelListeners) { processMultiMethodListeners(classLevelListeners, multiMethods, bean, beanName); // 然后跟着这行代码往下面这个方法看,接下来对取到的有@KafkaListenner类和方法进行了什么处理 } } return bean; } private void processMultiMethodListeners(Collection<KafkaListener> classLevelListeners, List<Method> multiMethods, Object bean, String beanName) { List<Method> checkedMethods = new ArrayList<Method>(); for (Method method : multiMethods) { checkedMethods.add(checkProxy(method, bean)); }
// 这里会遍历拿到的 classLevelListeners 集合,即每一个 kafkaListenner, 然后创建了一个 endpoint 对象,接下来看 processListenner 方法 for (KafkaListener classLevelListener : classLevelListeners) { MultiMethodKafkaListenerEndpoint<K, V> endpoint = new MultiMethodKafkaListenerEndpoint<K, V>(checkedMethods, bean); endpoint.setBeanFactory(this.beanFactory); processListener(endpoint, classLevelListener, bean, bean.getClass(), beanName); } } protected void processListener(MethodKafkaListenerEndpoint<?, ?> endpoint, KafkaListener kafkaListener, Object bean, Object adminTarget, String beanName) {
// 从这个方法可以知道,spring 会把拿到的 kafkaListenner 的一些信息(topic, partitions等)封装到那个 endpoint 里面 endpoint.setBean(bean); endpoint.setMessageHandlerMethodFactory(this.messageHandlerMethodFactory); endpoint.setId(getEndpointId(kafkaListener)); endpoint.setTopicPartitions(resolveTopicPartitions(kafkaListener)); endpoint.setTopics(resolveTopics(kafkaListener)); endpoint.setTopicPattern(resolvePattern(kafkaListener)); String group = kafkaListener.group(); if (StringUtils.hasText(group)) { Object resolvedGroup = resolveExpression(group); if (resolvedGroup instanceof String) { endpoint.setGroup((String) resolvedGroup); } } KafkaListenerContainerFactory<?> factory = null; String containerFactoryBeanName = resolve(kafkaListener.containerFactory()); if (StringUtils.hasText(containerFactoryBeanName)) { Assert.state(this.beanFactory != null, "BeanFactory must be set to obtain container factory by bean name"); try { factory = this.beanFactory.getBean(containerFactoryBeanName, KafkaListenerContainerFactory.class); } catch (NoSuchBeanDefinitionException ex) { throw new BeanInitializationException("Could not register Kafka listener endpoint on [" + adminTarget + "] for bean " + beanName + ", no " + KafkaListenerContainerFactory.class.getSimpleName() + " with id '" + containerFactoryBeanName + "' was found in the application context", ex); } } // 封装成了 endpoint 对象之后则进行注册 this.registrar.registerEndpoint(endpoint, factory); } }
通过分析这个 BPP 可以获得的信息是:在 IOC 容器初始化的时候获取所有带有 @KafkaListenner 注解的方法或类,然后将其信息封装到 endpoint 对象中,最后则对 endpoint 对象进行注册。接下来对注册进行分析 ↓
(2)KafkaListenerEndpointRegistrar 一个实现了 InitializingBean 的类,为啥要实现它呢?暂且埋下伏笔。
public class KafkaListenerEndpointRegistrar implements BeanFactoryAware, InitializingBean { private final List<KafkaListenerEndpointDescriptor> endpointDescriptors = new ArrayList<>(); public void registerEndpoint(KafkaListenerEndpoint endpoint, KafkaListenerContainerFactory<?> factory) { Assert.notNull(endpoint, "Endpoint must be set"); Assert.hasText(endpoint.getId(), "Endpoint id must be set"); // 可以看出 这里又将 endpoint 封装到了 KafkaListenerEndpointDescriptor 对象中,然后将其注册(保存)到了endpointDescriptors (一个List)中去。
// 于是从一开始根据 @KafkaListener 注解获取对象到注册则全部结束,是不是 consumer 创建就到此结束?当然还有一些工作,接下来就是真正核心的东西:创建 kafkaListener 容器 KafkaListenerEndpointDescriptor descriptor = new KafkaListenerEndpointDescriptor(endpoint, factory); synchronized (this.endpointDescriptors) { if (this.startImmediately) { // Register and start immediately this.endpointRegistry.registerListenerContainer(descriptor.endpoint, resolveContainerFactory(descriptor), true); } else { this.endpointDescriptors.add(descriptor); } } }
// 创建容器 start, 这就是为什么要实现 InitializingBean 接口的原因。因为在获取 @KafkaListener 的一些信息以及注册完成之后,还需要创建 KafkaListener 容器 @Override public void afterPropertiesSet() { registerAllEndpoints(); } protected void registerAllEndpoints() { synchronized (this.endpointDescriptors) {
// 这里发现,spring 对每一个 kafkaListener 注册一个 ListenerContainer 容器,即有多少个 @KafkaListener 它就会注册多少个 Listener 容器 for (KafkaListenerEndpointDescriptor descriptor : this.endpointDescriptors) { this.endpointRegistry.registerListenerContainer( descriptor.endpoint, resolveContainerFactory(descriptor)); } this.startImmediately = true; // trigger immediate startup } } }
从以上解析可以知道:spring 会将 @KafkaListener 注解中的信息注册到一个 list 集合中,信息注册完成之后则会进行容器的创建,有多少个 @KafkaListener 就注册多少个 Listener 容器。然而,容器到底是什么容器呢?是如何注册的?又有什么用呢?
(3)KafkaListenerEndpointRegistry 在这个类里我们将看到 Spring 是如何进行容器注册的
public class KafkaListenerEndpointRegistry implements DisposableBean, SmartLifecycle, ApplicationContextAware, ApplicationListener<ContextRefreshedEvent> { private final Map<String, MessageListenerContainer> listenerContainers = new ConcurrentHashMap<String, MessageListenerContainer>(); public void registerListenerContainer(KafkaListenerEndpoint endpoint, KafkaListenerContainerFactory<?> factory) { registerListenerContainer(endpoint, factory, false); }
public void registerListenerContainer(KafkaListenerEndpoint endpoint, KafkaListenerContainerFactory<?> factory, boolean startImmediately) { Assert.notNull(endpoint, "Endpoint must not be null"); Assert.notNull(factory, "Factory must not be null"); // 获取 id 即为 @KafkaListener 里的 id 作为 Map 的 key 值 String id = endpoint.getId(); Assert.hasText(id, "Endpoint id must not be empty"); synchronized (this.listenerContainers) { Assert.state(!this.listenerContainers.containsKey(id), "Another endpoint is already registered with id '" + id + "'");
// 创建容器,这里是创建什么容器,又是如何创建的呢? MessageListenerContainer container = createListenerContainer(endpoint, factory);
// 将容器注册到 listenerContainers (一个 Map)中 this.listenerContainers.put(id, container); if (StringUtils.hasText(endpoint.getGroup()) && this.applicationContext != null) { List<MessageListenerContainer> containerGroup; if (this.applicationContext.containsBean(endpoint.getGroup())) { containerGroup = this.applicationContext.getBean(endpoint.getGroup(), List.class); } else { containerGroup = new ArrayList<MessageListenerContainer>(); this.applicationContext.getBeanFactory().registerSingleton(endpoint.getGroup(), containerGroup); } containerGroup.add(container); }
// 容器启动 if (startImmediately) { startIfNecessary(container); } } } }
单单这个类远远满足不了我们对容器注册的好奇心,这里我们只能获取到这些信息:创建容器然后将容器注册到了一个 Map 中,最后是启动容器。然而容器是什么容器又是如何创建的还未知,于是接着摸索 ↓
(4)ConcurrentKafkaListenerContainerFactory 通过不懈的追踪,终于找到了 ‘元凶’.是的就是这个类创建了 Listener 容器。
public class ConcurrentKafkaListenerContainerFactory<K, V> extends AbstractKafkaListenerContainerFactory<ConcurrentMessageListenerContainer<K, V>, K, V> {
@Override protected ConcurrentMessageListenerContainer<K, V> createContainerInstance(KafkaListenerEndpoint endpoint) { Collection<TopicPartitionInitialOffset> topicPartitions = endpoint.getTopicPartitions(); if (!topicPartitions.isEmpty()) { ContainerProperties properties = new ContainerProperties( topicPartitions.toArray(new TopicPartitionInitialOffset[topicPartitions.size()])); return new ConcurrentMessageListenerContainer<K, V>(getConsumerFactory(), properties); } else { Collection<String> topics = endpoint.getTopics(); if (!topics.isEmpty()) { ContainerProperties properties = new ContainerProperties(topics.toArray(new String[topics.size()])); return new ConcurrentMessageListenerContainer<K, V>(getConsumerFactory(), properties); } else { ContainerProperties properties = new ContainerProperties(endpoint.getTopicPattern()); return new ConcurrentMessageListenerContainer<K, V>(getConsumerFactory(), properties); } } } }
哎呀,原来是创建了 ConcurrentMessageListenerContainer 啊,现在终于明白了 Spring 原来是创建这个 Listener 容器。然而这个容器又是何方圣神,come on ↓
(5)ConcurrentMessageListenerContainer 于是明白了容器的注册过程(创建 ConcurrentMessageListenerContainer 容器,然后把容器注册到 listenerContainers (一个 Map)中,最后就是启动容器)之后,就是研究容器究竟做了什么。
// 大家留意一下 AbstractMessageListenerContainer 这个类喔,待会有惊喜
public class ConcurrentMessageListenerContainer<K, V> extends AbstractMessageListenerContainer<K, V> { private final List<KafkaMessageListenerContainer<K, V>> containers = new ArrayList<>(); private int concurrency = 1; @Override protected void doStart() { if (!isRunning()) { ContainerProperties containerProperties = getContainerProperties(); TopicPartitionInitialOffset[] topicPartitions = containerProperties.getTopicPartitions();
// 这里可以知道,分区数不能小于这个 concurrency,这个东西到底是什么呢?现在知道它是一个 int 属性而且默认值为1 if (topicPartitions != null && this.concurrency > topicPartitions.length) { this.logger.warn("When specific partitions are provided, the concurrency must be less than or " + "equal to the number of partitions; reduced from " + this.concurrency + " to " + topicPartitions.length); this.concurrency = topicPartitions.length; } setRunning(true); // 然后根据 concurrency 遍历创建 KafkaMessageListenerContainer,并启动这个容器,最后将其注册到 containers (一个 List)中 for (int i = 0; i < this.concurrency; i++) { KafkaMessageListenerContainer<K, V> container; if (topicPartitions == null) { container = new KafkaMessageListenerContainer<>(this.consumerFactory, containerProperties); } else { container = new KafkaMessageListenerContainer<>(this.consumerFactory, containerProperties, partitionSubset(containerProperties, i)); } if (getBeanName() != null) { container.setBeanName(getBeanName() + "-" + i); } if (getApplicationEventPublisher() != null) { container.setApplicationEventPublisher(getApplicationEventPublisher()); } container.setClientIdSuffix("-" + i); container.start(); this.containers.add(container); } } } }
看了这里是不是有些疑问?为啥容器里面又创建容器?而且这个容器创建的个数还可以使用 concurrency 指定。那么这个 KafkaMessageListenerContainer 又是何方圣神,它和 ConcurrentMessageListenerContainer 又有什么关系呢?带着这些疑问继续追寻。
(6)KafkaMessageListenerContainer
// 大家发现没,这个类也继承了 AbstractMessageListenerContainer,这一点可以知道,此类和 ConcurrentMessageListenerContainer 是来自一个老子的(兄弟关系)
public class KafkaMessageListenerContainer<K, V> extends AbstractMessageListenerContainer<K, V> { private GenericAcknowledgingMessageListener<?> acknowledgingMessageListener; @Override protected void doStart() { if (isRunning()) { return; } ContainerProperties containerProperties = getContainerProperties(); if (!this.consumerFactory.isAutoCommit()) { AckMode ackMode = containerProperties.getAckMode(); if (ackMode.equals(AckMode.COUNT) || ackMode.equals(AckMode.COUNT_TIME)) { Assert.state(containerProperties.getAckCount() > 0, "'ackCount' must be > 0"); } if ((ackMode.equals(AckMode.TIME) || ackMode.equals(AckMode.COUNT_TIME)) && containerProperties.getAckTime() == 0) { containerProperties.setAckTime(5000); } } Object messageListener = containerProperties.getMessageListener(); Assert.state(messageListener != null, "A MessageListener is required"); if (messageListener instanceof GenericAcknowledgingMessageListener) { this.acknowledgingMessageListener = (GenericAcknowledgingMessageListener<?>) messageListener; } else if (messageListener instanceof GenericMessageListener) { this.listener = (GenericMessageListener<?>) messageListener; } else { throw new IllegalStateException("messageListener must be 'MessageListener' " + "or 'AcknowledgingMessageListener', not " + messageListener.getClass().getName()); } if (containerProperties.getConsumerTaskExecutor() == null) { SimpleAsyncTaskExecutor consumerExecutor = new SimpleAsyncTaskExecutor( (getBeanName() == null ? "" : getBeanName()) + "-C-"); containerProperties.setConsumerTaskExecutor(consumerExecutor); } if (containerProperties.getListenerTaskExecutor() == null) { SimpleAsyncTaskExecutor listenerExecutor = new SimpleAsyncTaskExecutor( (getBeanName() == null ? "" : getBeanName()) + "-L-"); containerProperties.setListenerTaskExecutor(listenerExecutor); }
// 终于,在这里我们知道原来 Spring 是在这里创建的 kafka consumer,不信可以到这个 ListenerConsumer 里去瞧瞧 this.listenerConsumer = new ListenerConsumer(this.listener, this.acknowledgingMessageListener); setRunning(true);
// 看到这个 Future 似乎 get 到了啥东西。是的,多线程,这里使用了多线程-为每一个listenerConsumer开启一个线程 this.listenerConsumerFuture = containerProperties .getConsumerTaskExecutor() .submitListenable(this.listenerConsumer); } }
通过对这个类的分析,我们明白了 KafkaMessageListenerContainer 和 ConcurrentMessageListenerContainer 的关系,而且还发现 Spring 实际创建 kafka consumer 也是在这个容器中发生的,而且还嗅到了多线程。于是进一步确认,经过资料查阅果然:
//接口,封装原生KafkaConsumer,一个container封装一个consumer interface MessageListenerContainer; //单线程container实现,只启动一个consumer class KafkaMessageListenerContainer implemets MessageListenerContainer; //多线程container实现,负责创建多个KafkaMessageListenerContainer class ConcurrentMessageListenerContainer implemets MessageListenerContainer; //接口,工厂模式,container工厂,负责创建container,当使用@KafkaListener时需要提供 interface KafkaListenerContainerFactory<C extends MessageListenerContainer>; //container工厂的唯一实现,且参数为多线程container,如果需要单线程,setConsurrency(null)即可,这也是默认参数 class KafkaListenerContainerFactory<ConcurrentMessageListenerContainer<K, V>>
大家看到了没有,一切都水落石出:原来 Spring 提供了单线程和多线程两种方式使用 kafka consumer,而是否使用多线程由之前那个 int 属性 concurrency 变量决定,而且在分析 ConcurrentMessageListenerContainer 的时候知道它的默认值是1,也就是默认是使用单线程的。而且这个线程数还不能大于分区数。顺便提及一下:spring kafka 提供了两种提交方式,手动提交和自动提交,而手动提交 Spring 是如何监听到的呢,这个就是在 KafkaMessageListenerContainer 我们看到的那个 acknowledgingMessageListener 做出的贡献了。
消费者剩下花絮请看:spring kafka consumer原理解析二