This version uses the Apache Kafka 0.10.0.x or 0.10.1.x client.

Listeners can be configured to receive the entire batch of messages returned by the consumer.poll() operation, rather than one at a time.

Null payloads are used to "delete" keys when using log compaction.

When explicitly assigning partitions, you can now configure the initial offset relative to the current position for the consumer group, rather than absolute or relative to the current end.

You can now seek the position of each topic/partition. This can be used to set the initial position during initialization when group management is in use and Kafka assigns the partitions. You can also seek when an idle container is detected, or at any arbitrary point in your application’s execution. See the section called “Seeking to a Specific Offset” for more information.

This is the 5 minute tour to get started with Spring Kafka.

Prerequisites: install and run Apache Kafka Then grab the spring-kafka JAR and all of its dependencies - the easiest way to do that is to declare a dependency in your build tool, e.g. for Maven:

<dependency>
  <groupId>org.springframework.kafka</groupId>
  <artifactId>spring-kafka</artifactId>
  <version>1.2.0.RELEASE</version>
</dependency>

And for Gradle:

compile 'org.springframework.kafka:spring-kafka:1.2.0.RELEASE'

@Test
public void testAutoCommit() throws Exception {
    logger.info("Start auto");
    ContainerProperties containerProps = new ContainerProperties("topic1", "topic2");
    final CountDownLatch latch = new CountDownLatch(4);
    containerProps.setMessageListener(new MessageListener<Integer, String>() {

        @Override
        public void onMessage(ConsumerRecord<Integer, String> message) {
            logger.info("received: " + message);
            latch.countDown();
        }

    });
    KafkaMessageListenerContainer<Integer, String> container = createContainer(containerProps);
    container.setBeanName("testAuto");
    container.start();
    Thread.sleep(1000); // wait a bit for the container to start
    KafkaTemplate<Integer, String> template = createTemplate();
    template.setDefaultTopic(topic1);
    template.sendDefault(0, "foo");
    template.sendDefault(2, "bar");
    template.sendDefault(0, "baz");
    template.sendDefault(2, "qux");
    template.flush();
    assertTrue(latch.await(60, TimeUnit.SECONDS));
    container.stop();
    logger.info("Stop auto");

}

private KafkaMessageListenerContainer<Integer, String> createContainer(
                        ContainerProperties containerProps) {
    Map<String, Object> props = consumerProps();
    DefaultKafkaConsumerFactory<Integer, String> cf =
                            new DefaultKafkaConsumerFactory<Integer, String>(props);
    KafkaMessageListenerContainer<Integer, String> container =
                            new KafkaMessageListenerContainer<>(cf, containerProps);
    return container;
}

private KafkaTemplate<Integer, String> createTemplate() {
    Map<String, Object> senderProps = senderProps();
    ProducerFactory<Integer, String> pf =
              new DefaultKafkaProducerFactory<Integer, String>(senderProps);
    KafkaTemplate<Integer, String> template = new KafkaTemplate<>(pf);
    return template;
}

private Map<String, Object> consumerProps() {
    Map<String, Object> props = new HashMap<>();
    props.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, "localhost:9092");
    props.put(ConsumerConfig.GROUP_ID_CONFIG, group);
    props.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG, true);
    props.put(ConsumerConfig.AUTO_COMMIT_INTERVAL_MS_CONFIG, "100");
    props.put(ConsumerConfig.SESSION_TIMEOUT_MS_CONFIG, "15000");
    props.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, IntegerDeserializer.class);
    props.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class);
    return props;
}

private Map<String, Object> senderProps() {
    Map<String, Object> props = new HashMap<>();
    props.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, "localhost:9092");
    props.put(ProducerConfig.RETRIES_CONFIG, 0);
    props.put(ProducerConfig.BATCH_SIZE_CONFIG, 16384);
    props.put(ProducerConfig.LINGER_MS_CONFIG, 1);
    props.put(ProducerConfig.BUFFER_MEMORY_CONFIG, 33554432);
    props.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, IntegerSerializer.class);
    props.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, StringSerializer.class);
    return props;
}

@Autowired
private Listener listener;

@Autowired
private KafkaTemplate<Integer, String> template;

@Test
public void testSimple() throws Exception {
    template.send("annotated1", 0, "foo");
    template.flush();
    assertTrue(this.listener.latch1.await(10, TimeUnit.SECONDS));
}

@Configuration
@EnableKafka
public class Config {

    @Bean
    ConcurrentKafkaListenerContainerFactory<Integer, String>
                        kafkaListenerContainerFactory() {
        ConcurrentKafkaListenerContainerFactory<Integer, String> factory =
                                new ConcurrentKafkaListenerContainerFactory<>();
        factory.setConsumerFactory(consumerFactory());
        return factory;
    }

    @Bean
    public ConsumerFactory<Integer, String> consumerFactory() {
        return new DefaultKafkaConsumerFactory<>(consumerConfigs());
    }

    @Bean
    public Map<String, Object> consumerConfigs() {
        Map<String, Object> props = new HashMap<>();
        props.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, embeddedKafka.getBrokersAsString());
        ...
        return props;
    }

    @Bean
    public Listener listener() {
        return new Listener();
    }

    @Bean
    public ProducerFactory<Integer, String> producerFactory() {
        return new DefaultKafkaProducerFactory<>(producerConfigs());
    }

    @Bean
    public Map<String, Object> producerConfigs() {
        Map<String, Object> props = new HashMap<>();
        props.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, embeddedKafka.getBrokersAsString());
        ...
        return props;
    }

    @Bean
    public KafkaTemplate<Integer, String> kafkaTemplate() {
        return new KafkaTemplate<Integer, String>(producerFactory());
    }

}

public class Listener {

    private final CountDownLatch latch1 = new CountDownLatch(1);

    @KafkaListener(id = "foo", topics = "annotated1")
    public void listen1(String foo) {
        this.latch1.countDown();
    }

}

@SpringBootApplication
public class Application implements CommandLineRunner {

    public static Logger logger = LoggerFactory.getLogger(Application.class);

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

    @Autowired
    private KafkaTemplate<String, String> template;

    private final CountDownLatch latch = new CountDownLatch(3);

    @Override
    public void run(String... args) throws Exception {
        this.template.send("myTopic", "foo1");
        this.template.send("myTopic", "foo2");
        this.template.send("myTopic", "foo3");
        latch.await(60, TimeUnit.SECONDS);
        logger.info("All received");
    }

    @KafkaListener(topics = "myTopic")
    public void listen(ConsumerRecord<?, ?> cr) throws Exception {
        logger.info(cr.toString());
        latch.countDown();
    }

}

spring.kafka.consumer.group-id=foo
spring.kafka.consumer.auto-offset-reset=earliest

Messages can be received by configuring a MessageListenerContainer and providing a Message Listener, or by using the @KafkaListener annotation.

public interface MessageListener<K, V> {} 

    void onMessage(ConsumerRecord<K, V> data);

}

public interface AcknowledgingMessageListener<K, V> {} 

    void onMessage(ConsumerRecord<K, V> data, Acknowledgment acknowledgment);

}

public interface BatchMessageListener<K, V> {} 

    void onMessage(List<ConsumerRecord<K, V>> data);

}

public interface BatchAcknowledgingMessageListener<K, V> {} 

    void onMessage(List<ConsumerRecord<K, V>> data, Acknowledgment acknowledgment);

}

Message Listener Containers

Two MessageListenerContainer implementations are provided:

• KafkaMessageListenerContainer
• ConcurrentMessageListenerContainer

The KafkaMessageListenerContainer receives all message from all topics/partitions on a single thread. The ConcurrentMessageListenerContainer delegates to 1 or more KafkaMessageListenerContainer s to provide multi-threaded consumption.

KafkaMessageListenerContainer

The following constructors are available.

public KafkaMessageListenerContainer(ConsumerFactory<K, V> consumerFactory,
                    ContainerProperties containerProperties)

public KafkaMessageListenerContainer(ConsumerFactory<K, V> consumerFactory,
                    ContainerProperties containerProperties,
                    TopicPartitionInitialOffset... topicPartitions)

Each takes a ConsumerFactory and information about topics and partitions, as well as other configuration in a ContainerProperties object. The second constructor is used by the ConcurrentMessageListenerContainer (see below) to distribute TopicPartitionInitialOffset across the consumer instances. ContainerProperties has the following constructors:

public ContainerProperties(TopicPartitionInitialOffset... topicPartitions)

public ContainerProperties(String... topics)

public ContainerProperties(Pattern topicPattern)

The first takes an array of TopicPartitionInitialOffset arguments to explicitly instruct the container which partitions to use (using the consumer assign() method), and with an optional initial offset: a positive value is an absolute offset by default; a negative value is relative to the current last offset within a partition by default. A constructor for TopicPartitionInitialOffset is provided that takes an additional boolean argument. If this is true, the initial offsets (positive or negative) are relative to the current position for this consumer. The offsets are applied when the container is started. The second takes an array of topics and Kafka allocates the partitions based on the group.id property - distributing partitions across the group. The third uses a regex Pattern to select the topics.

Refer to the JavaDocs for ContainerProperties for more information about the various properties that can be set.

ConcurrentMessageListenerContainer

The single constructor is similar to the first KafkaListenerContainer constructor:

public ConcurrentMessageListenerContainer(ConsumerFactory<K, V> consumerFactory,
                            ContainerProperties containerProperties)

It also has a property concurrency, e.g. container.setConcurrency(3) will create 3 KafkaMessageListenerContainer s.

For the first constructor, kafka will distribute the partitions across the consumers. For the second constructor, the ConcurrentMessageListenerContainer distributes the TopicPartition s across the delegate KafkaMessageListenerContainer s.

If, say, 6 TopicPartition s are provided and the concurrency is 3; each container will get 2 partitions. For 5 TopicPartition s, 2 containers will get 2 partitions and the third will get 1. If the concurrency is greater than the number of TopicPartitions, the concurrency will be adjusted down such that each container will get one partition.

Committing Offsets

Several options are provided for committing offsets. If the enable.auto.commit consumer property is true, kafka will auto-commit the offsets according to its configuration. If it is false, the containers support the following AckMode s.

The consumer poll() method will return one or more ConsumerRecords; the MessageListener is called for each record; the following describes the action taken by the container for each AckMode :

• RECORD - commit the offset when the listener returns after processing the record.
• BATCH - commit the offset when all the records returned by the poll() have been processed.
• TIME - commit the offset when all the records returned by the poll() have been processed as long as the ackTime since the last commit has been exceeded.
• COUNT - commit the offset when all the records returned by the poll() have been processed as long as ackCount records have been received since the last commit.
• COUNT_TIME - similar to TIME and COUNT but the commit is performed if either condition is true.
• MANUAL - the message listener is responsible to acknowledge() the Acknowledgment; after which, the same semantics as BATCH are applied.
• MANUAL_IMMEDIATE - commit the offset immediately when the Acknowledgment.acknowledge() method is called by the listener.

	Note
	MANUAL, and MANUAL_IMMEDIATE require the listener to be an AcknowledgingMessageListener or a BatchAcknowledgingMessageListener; see Message Listeners.

The commitSync() or commitAsync() method on the consumer is used, depending on the syncCommits container property.

The Acknowledgment has this method:

public interface Acknowledgment {

    void acknowledge();

}

This gives the listener control over when offsets are committed.

public class Listener {

    @KafkaListener(id = "foo", topics = "myTopic")
    public void listen(String data) {
        ...
    }

}

@Configuration
@EnableKafka
public class KafkaConfig {

    @Bean
    KafkaListenerContainerFactory<ConcurrentMessageListenerContainer<Integer, String>>
                        kafkaListenerContainerFactory() {
        ConcurrentKafkaListenerContainerFactory<Integer, String> factory =
                                new ConcurrentKafkaListenerContainerFactory<>();
        factory.setConsumerFactory(consumerFactory());
        factory.setConcurrency(3);
        factory.getContainerProperties().setPollTimeout(3000);
        return factory;
    }

    @Bean
    public ConsumerFactory<Integer, String> consumerFactory() {
        return new DefaultKafkaConsumerFactory<>(consumerConfigs());
    }

    @Bean
    public Map<String, Object> consumerConfigs() {
        Map<String, Object> props = new HashMap<>();
        props.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, embeddedKafka.getBrokersAsString());
        ...
        return props;
    }
}

@KafkaListener(id = "bar", topicPartitions =
        { @TopicPartition(topic = "topic1", partitions = { "0", "1" }),
          @TopicPartition(topic = "topic2", partitions = "0",
             partitionOffsets = @PartitionOffset(partition = "1", initialOffset = "100"))
        })
public void listen(ConsumerRecord<?, ?> record) {
    ...
}

@KafkaListener(id = "baz", topics = "myTopic",
          containerFactory = "kafkaManualAckListenerContainerFactory")
public void listen(String data, Acknowledgment ack) {
    ...
    ack.acknowledge();
}

@KafkaListener(id = "qux", topicPattern = "myTopic1")
public void listen(@Payload String foo,
        @Header(KafkaHeaders.RECEIVED_MESSAGE_KEY) Integer key,
        @Header(KafkaHeaders.RECEIVED_PARTITION_ID) int partition,
        @Header(KafkaHeaders.RECEIVED_TOPIC) String topic) {
    ...
}

@Bean
public KafkaListenerContainerFactory<?> batchFactory() {
    ConcurrentKafkaListenerContainerFactory<Integer, String> factory =
            new ConcurrentKafkaListenerContainerFactory<>();
    factory.setConsumerFactory(consumerFactory());
    factory.setBatchListener(true);  // <<<<<<<<<<<<<<<<<<<<<<<<<
    return factory;
}

@KafkaListener(id = "list", topics = "myTopic", containerFactory = "batchFactory")
public void listen(List<String> list) {
    ...
}

@KafkaListener(id = "list", topics = "myTopic", containerFactory = "batchFactory")
public void listen(List<String> list,
        @Header(KafkaHeaders.RECEIVED_MESSAGE_KEY) List<Integer> keys,
        @Header(KafkaHeaders.RECEIVED_PARTITION_ID) List<Integer> partitions,
        @Header(KafkaHeaders.RECEIVED_TOPIC) List<String> topics,
        @Header(KafkaHeaders.OFFSET) List<Long> offsets) {
    ...
}

@KafkaListener(id = "listMsg", topics = "myTopic", containerFactory = "batchFactory")
public void listen14(List<Message<?>> list) {
    ...
}

@KafkaListener(id = "listMsgAck", topics = "myTopic", containerFactory = "batchFactory")
public void listen15(List<Message<?>> list, Acknowledgment ack) {
    ...
}

@KafkaListener(id = "listCRs", topics = "myTopic", containerFactory = "batchFactory")
public void listen(List<ConsumerRecord<Integer, String>> list) {
    ...
}

@KafkaListener(id = "listCRsAck", topics = "myTopic", containerFactory = "batchFactory")
public void listen(List<ConsumerRecord<Integer, String>> list, Acknowledgment ack) {
    ...
}

Retrying Deliveries

If your listener throws an exception, the default behavior is to invoke the ErrorHandler, if configured, or logged otherwise.

	Note
	Two error handler interfaces are provided ErrorHandler and BatchErrorHandler; the appropriate type must be configured to match the Message Listener.

To retry deliveries, convenient listener adapters - RetryingMessageListenerAdapter and RetryingAcknowledgingMessageListenerAdapter are provided, depending on whether you are using a MessageListener or an AcknowledgingMessageListener.

These can be configured with a RetryTemplate and RecoveryCallback<Void> - see the spring-retry project for information about these components. If a recovery callback is not provided, the exception is thrown to the container after retries are exhausted. In that case, the ErrorHandler will be invoked, if configured, or logged otherwise.

When using @KafkaListener, set the RetryTemplate (and optionally recoveryCallback) on the container factory and the listener will be wrapped in the appropriate retrying adapter.

A retry adapter is not provided for any of the batch message listeners.

Detecting Idle Asynchronous Consumers

While efficient, one problem with asynchronous consumers is detecting when they are idle - users might want to take some action if no messages arrive for some period of time.

You can configure the listener container to publish a ListenerContainerIdleEvent when some time passes with no message delivery. While the container is idle, an event will be published every idleEventInterval milliseconds.

To configure this feature, set the idleEventInterval on the container:

@Bean
public KafKaMessageListenerContainer(ConnectionFactory connectionFactory) {
    ContainerProperties containerProps = new ContainerProperties("topic1", "topic2");
    ...
    containerProps.setIdleEventInterval(60000L);
    ...
    KafKaMessageListenerContainer<String, String> container = new KafKaMessageListenerContainer<>(...);
    return container;
}

Or, for a @KafkaListener…​

@Bean
public ConcurrentKafkaListenerContainerFactory kafkaListenerContainerFactory() {
    ConcurrentKafkaListenerContainerFactory<String, String> factory =
                new ConcurrentKafkaListenerContainerFactory<>();
    ...
    factory.getContainerProperties().setIdleEventInterval(60000L);
    ...
    return factory;
}

In each of these cases, an event will be published once per minute while the container is idle.

Event Consumption

You can capture these events by implementing ApplicationListener - either a general listener, or one narrowed to only receive this specific event. You can also use @EventListener, introduced in Spring Framework 4.2.

The following example combines the @KafkaListener and @EventListener into a single class. It’s important to understand that the application listener will get events for all containers so you may need to check the listener id if you want to take specific action based on which container is idle. You can also use the @EventListener condition for this purpose.

The events have 4 properties:

• source - the listener container instance
• id - the listener id (or container bean name)
• idleTime - the time the container had been idle when the event was published
• topicPartitions - the topics/partitions that the container was assigned at the time the event was generated

public class Listener {

    @KafkaListener(id = "qux", topics = "annotated")
    public void listen4(@Payload String foo, Acknowledgment ack) {
        ...
    }

    @EventListener(condition = "event.listenerId.startsWith('qux-')")
    public void eventHandler(ListenerContainerIdleEvent event) {
        this.event = event;
        eventLatch.countDown();
    }

}

	Important
	Event listeners will see events for all containers; so, in the example above, we narrow the events received based on the listener ID. Since containers created for the @KafkaListener support concurrency, the actual containers are named id-n where the n is a unique value for each instance to support the concurrency. Hence we use startsWith in the condition.

Caution

If you wish to use the idle event to stop the lister container, you should not call container.stop() on the thread that calls the listener - it will cause delays and unnecessary log messages. Instead, you should hand off the event to a different thread that can then stop the container. Also, you should not stop() the container instance in the event if it is a child container, you should stop the concurrent container instead.

Current Positions when Idle

Note that you can obtain the current positions when idle is detected by implementing ConsumerSeekAware in your listener; see onIdleContainer() in `the section called “Seeking to a Specific Offset”.

void registerSeekCallback(ConsumerSeekCallback callback);

void onPartitionsAssigned(Map<TopicPartition, Long> assignments, ConsumerSeekCallback callback);

void onIdleContainer(Map<TopicPartition, Long> assignments, ConsumerSeekCallback callback);

void seek(String topic, int partition, long offset);

Apache Kafka provides a high-level API for serializing/deserializing record values as well as their keys. It is present with the org.apache.kafka.common.serialization.Serializer<T> and org.apache.kafka.common.serialization.Deserializer<T> abstractions with some built-in implementations. Meanwhile we can specify simple (de)serializer classes using Producer and/or Consumer configuration properties, e.g.:

props.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, IntegerDeserializer.class);
props.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class);
...
props.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, IntegerSerializer.class);
props.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, StringSerializer.class);

for more complex or particular cases, the KafkaConsumer, and therefore KafkaProducer, provides overloaded constructors to accept (De)Serializer instances for keys and/or values, respectively.

To meet this API, the DefaultKafkaProducerFactory and DefaultKafkaConsumerFactory also provide properties to allow to inject a custom (De)Serializer to target Producer/Consumer.

For this purpose, Spring for Apache Kafka also provides JsonSerializer/JsonDeserializer implementations based on the Jackson JSON object mapper. The JsonSerializer is quite simple and just allows writing any Java object as a JSON byte[], the JsonDeserializer requires an additional Class<?> targetType argument to allow the deserialization of a consumed byte[] to the proper target object.

JsonDeserializer<Bar> barDeserializer = new JsonDeserializer<>(Bar.class);

Both JsonSerializer and JsonDeserializer can be customized with an ObjectMapper. You can also extend them to implement some particular configuration logic in the configure(Map<String, ?> configs, boolean isKey) method.

Although the Serializer/Deserializer API is quite simple and flexible from the low-level Kafka Consumer and Producer perspective, you might need more flexibility at the Spring Messaging level, either when using @KafkaListener or Spring Integration. To easily convert to/from org.springframework.messaging.Message, Spring for Apache Kafka provides a MessageConverter abstraction with the MessagingMessageConverter implementation and its StringJsonMessageConverter customization. The MessageConverter can be injected into KafkaTemplate instance directly and via AbstractKafkaListenerContainerFactory bean definition for the @KafkaListener.containerFactory() property:

@Bean
public KafkaListenerContainerFactory<?> kafkaJsonListenerContainerFactory() {
    ConcurrentKafkaListenerContainerFactory<Integer, String> factory =
        new ConcurrentKafkaListenerContainerFactory<>();
    factory.setConsumerFactory(consumerFactory());
    factory.setMessageConverter(new StringJsonMessageConverter());
    return factory;
}
...
@KafkaListener(topics = "jsonData",
                containerFactory = "kafkaJsonListenerContainerFactory")
public void jsonListener(Foo foo) {
...
}

When using a @KafkaListener, the parameter type is provided to the message converter to assist with the conversion.

	Note
	When using the StringJsonMessageConverter, you should use a StringDeserializer in the kafka consumer configuration and StringSerializer in the kafka producer configuration, when using Spring Integration or the KafkaTemplate.send(Message<?> message) method.

When using Log Compaction, it is possible to send and receive messages with null payloads which identifies the deletion of a key.

Starting with version 1.0.3, this is now fully supported.

To send a null payload using the KafkaTemplate simply pass null into the value argument of the send() methods. One exception to this is the send(Message<?> message) variant. Since spring-messaging Message<?> cannot have a null payload, a special payload type KafkaNull is used and the framework will send null. For convenience, the static KafkaNull.INSTANCE is provided.

When using a message listener container, the received ConsumerRecord will have a null value().

To configure the @KafkaListener to handle null payloads, you must use the @Payload annotation with required = false; you will usually also need the key so your application knows which key was "deleted":

@KafkaListener(id = "deletableListener", topics = "myTopic")
public void listen(@Payload(required = false) String value, @Header(KafkaHeaders.RECEIVED_MESSAGE_KEY) String key) {
    // value == null represents key deletion
}

When using a class-level @KafkaListener, some additional configuration is needed - a @KafkaHandler method with a KafkaNull payload:

@KafkaListener(id = "multi", topics = "myTopic")
static class MultiListenerBean {

    private final CountDownLatch latch1 = new CountDownLatch(2);

    @KafkaHandler
    public void listen(String foo) {
        ...
    }

    @KafkaHandler
    public void listen(Integer bar) {
        ...
    }

    @KafkaHandler
    public void delete(@Payload(required = false) KafkaNull nul, @Header(KafkaHeaders.RECEIVED_MESSAGE_KEY) int key) {
        ...
    }

}

Starting with version 1.1.4, Spring for Apache Kafka provides first class support for Kafka Streams. For using it from a Spring application, the kafka-streams jar must be present on classpath. It is an optional dependency of the spring-kafka project and isn’t downloaded transitively.

The reference Apache Kafka Streams documentation suggests this way of using the API:

// Use the builders to define the actual processing topology, e.g. to specify
// from which input topics to read, which stream operations (filter, map, etc.)
// should be called, and so on.

KStreamBuilder builder = ...;  // when using the Kafka Streams DSL
//
// OR
//
TopologyBuilder builder = ...; // when using the Processor API

// Use the configuration to tell your application where the Kafka cluster is,
// which serializers/deserializers to use by default, to specify security settings,
// and so on.
StreamsConfig config = ...;

KafkaStreams streams = new KafkaStreams(builder, config);

// Start the Kafka Streams instance
streams.start();

// Stop the Kafka Streams instance
streams.close();

So, we have two main components: KStreamBuilder (which extends TopologyBuilder as well) with an API to build KStream (or KTable) instances and KafkaStreams to manage their lifecycle. Note: all KStream instances exposed to a KafkaStreams instance by a single KStreamBuilder will be started and stopped at the same time, even if they have a fully different logic. In other words all our streams defined by a KStreamBuilder are tied with a single lifecycle control. Once a KafkaStreams instance has been closed via streams.close() it cannot be restarted, and a new KafkaStreams instance to restart stream processing must be created instead.

To simplify the usage of Kafka Streams from the Spring application context perspective and utilize the lifecycle management via container, the Spring for Apache Kafka introduces KStreamBuilderFactoryBean. This is an AbstractFactoryBean implementation to expose a KStreamBuilder singleton instance as a bean:

@Bean
public FactoryBean<KStreamBuilder> myKStreamBuilder(StreamsConfig streamsConfig) {
    return new KStreamBuilderFactoryBean(streamsConfig);
}

The KStreamBuilderFactoryBean also implements SmartLifecycle to manage lifecycle of an internal KafkaStreams instance. Similar to the Kafka Streams API, the KStream instances must be defined before starting the KafkaStreams, and that also applies for the Spring API for Kafka Streams. Therefore we have to declare KStream s on the KStreamBuilder before the application context is refreshed, when we use default autoStartup = true on the KStreamBuilderFactoryBean. For example, KStream can be just as a regular bean definition, meanwhile the Kafka Streams API is used without any impacts:

@Bean
public KStream<?, ?> kStream(KStreamBuilder kStreamBuilder) {
    KStream<Integer, String> stream = kStreamBuilder.stream(STREAMING_TOPIC1);
    // Fluent KStream API
    return stream;
}

If you would like to control lifecycle manually (e.g. stop and start by some condition), you can reference the KStreamBuilderFactoryBean bean directly using factory bean (&) prefix. Since KStreamBuilderFactoryBean utilize its internal KafkaStreams instance, it is safe to stop and restart it again - a new KafkaStreams is created on each start(). Also consider using different KStreamBuilderFactoryBean s, if you would like to control lifecycles for KStream instances separately.

To configure the Kafka Streams environment, the KStreamBuilderFactoryBean requires a Map of particular properties or a StreamsConfig instance. See Apache Kafka documentation for all possible options.

To avoid boilerplate code for most cases, especially when you develop micro services, Spring for Apache Kafka provides the @EnableKafkaStreams annotation, which should be placed alongside with @Configuration. Only you need is to declare StreamsConfig bean with the defaultKafkaStreamsConfig name. A KStreamBuilder bean with the defaultKStreamBuilder name will be declare in the application context automatically. Any additional KStreamBuilderFactoryBean beans can be declared and used as well.

Putting it all together:

@Configuration
@EnableKafka
@EnableKafkaStreams
public static class KafkaStreamsConfiguration {

    @Bean(name = KafkaStreamsDefaultConfiguration.DEFAULT_STREAMS_CONFIG_BEAN_NAME)
    public StreamsConfig kStreamsConfigs() {
        Map<String, Object> props = new HashMap<>();
        props.put(StreamsConfig.APPLICATION_ID_CONFIG, "testStreams");
        props.put(StreamsConfig.KEY_SERDE_CLASS_CONFIG, Serdes.Integer().getClass().getName());
        props.put(StreamsConfig.VALUE_SERDE_CLASS_CONFIG, Serdes.String().getClass().getName());
        props.put(StreamsConfig.TIMESTAMP_EXTRACTOR_CLASS_CONFIG, WallclockTimestampExtractor.class.getName());
        return new StreamsConfig(props);
    }

    @Bean
    public KStream<Integer, String> kStream(KStreamBuilder kStreamBuilder) {
        KStream<Integer, String> stream = kStreamBuilder.stream("streamingTopic1");
        stream
                .mapValues(String::toUpperCase)
                .groupByKey()
                .reduce((String value1, String value2) -> value1 + value2,
                		TimeWindows.of(1000),
                		"windowStore")
                .toStream()
                .map((windowedId, value) -> new KeyValue<>(windowedId.key(), value))
                .filter((i, s) -> s.length() > 40)
                .to("streamingTopic2");

        stream.print();

        return stream;
    }

}

The spring-kafka-test jar contains some useful utilities to assist with testing your applications.

o.s.kafka.test.utils.KafkaUtils provides some static methods to set up producer and consumer properties:

/**
 * Set up test properties for an {@code <Integer, String>} consumer.
 * @param group the group id.
 * @param autoCommit the auto commit.
 * @param embeddedKafka a {@link KafkaEmbedded} instance.
 * @return the properties.
 */
public static Map<String, Object> consumerProps(String group, String autoCommit,
                                       KafkaEmbedded embeddedKafka) { ... }

/**
 * Set up test properties for an {@code <Integer, String>} producer.
 * @param embeddedKafka a {@link KafkaEmbedded} instance.
 * @return the properties.
 */
public static Map<String, Object> senderProps(KafkaEmbedded embeddedKafka) { ... }

A JUnit @Rule is provided that creates an embedded kafka server.

/**
 * Create embedded Kafka brokers.
 * @param count the number of brokers.
 * @param controlledShutdown passed into TestUtils.createBrokerConfig.
 * @param topics the topics to create (2 partitions per).
 */
public KafkaEmbedded(int count, boolean controlledShutdown, String... topics) { ... }

/**
 *
 * Create embedded Kafka brokers.
 * @param count the number of brokers.
 * @param controlledShutdown passed into TestUtils.createBrokerConfig.
 * @param partitions partitions per topic.
 * @param topics the topics to create.
 */
public KafkaEmbedded(int count, boolean controlledShutdown, int partitions, String... topics) { ... }

The embedded kafka class has a utility method allowing you to consume for all the topics it created:

Map<String, Object> consumerProps = KafkaTestUtils.consumerProps("testT", "false", embeddedKafka);
DefaultKafkaConsumerFactory<Integer, String> cf = new DefaultKafkaConsumerFactory<Integer, String>(
        consumerProps);
Consumer<Integer, String> consumer = cf.createConsumer();
embeddedKafka.consumeFromAllEmbeddedTopics(consumer);

The KafkaTestUtils has some utility methods to fetch results from the consumer:

/**
 * Poll the consumer, expecting a single record for the specified topic.
 * @param consumer the consumer.
 * @param topic the topic.
 * @return the record.
 * @throws org.junit.ComparisonFailure if exactly one record is not received.
 */
public static <K, V> ConsumerRecord<K, V> getSingleRecord(Consumer<K, V> consumer, String topic) { ... }

/**
 * Poll the consumer for records.
 * @param consumer the consumer.
 * @return the records.
 */
public static <K, V> ConsumerRecords<K, V> getRecords(Consumer<K, V> consumer) { ... }

Usage:

...
template.sendDefault(0, 2, "bar");
ConsumerRecord<Integer, String> received = KafkaTestUtils.getSingleRecord(consumer, "topic");
...

When the embedded server is started by JUnit, it sets a system property spring.embedded.kafka.brokers to the address of the broker(s). A convenient constant KafkaEmbedded.SPRING_EMBEDDED_KAFKA_BROKERS is provided for this property.

The o.s.kafka.test.hamcrest.KafkaMatchers provides the following matchers:

/**
 * @param key the key
 * @param <K> the type.
 * @return a Matcher that matches the key in a consumer record.
 */
public static <K> Matcher<ConsumerRecord<K, ?>> hasKey(K key) { ... }

/**
 * @param value the value.
 * @param <V> the type.
 * @return a Matcher that matches the value in a consumer record.
 */
public static <V> Matcher<ConsumerRecord<?, V>> hasValue(V value) { ... }

/**
 * @param partition the partition.
 * @return a Matcher that matches the partition in a consumer record.
 */
public static Matcher<ConsumerRecord<?, ?>> hasPartition(int partition) { ... }

/**
 * @param key the key
 * @param <K> the type.
 * @return a Condition that matches the key in a consumer record.
 */
public static <K> Condition<ConsumerRecord<K, ?>> key(K key) { ... }

/**
 * @param value the value.
 * @param <V> the type.
 * @return a Condition that matches the value in a consumer record.
 */
public static <V> Condition<ConsumerRecord<?, V>> value(V value) { ... }

/**
 * @param partition the partition.
 * @return a Condition that matches the partition in a consumer record.
 */
public static Condition<ConsumerRecord<?, ?>> partition(int partition) { ... }

Putting it all together:

public class KafkaTemplateTests {

    private static final String TEMPLATE_TOPIC = "templateTopic";

    @ClassRule
    public static KafkaEmbedded embeddedKafka = new KafkaEmbedded(1, true, TEMPLATE_TOPIC);

    @Test
    public void testTemplate() throws Exception {
        Map<String, Object> consumerProps = KafkaTestUtils.consumerProps("testT", "false", embeddedKafka);
        DefaultKafkaConsumerFactory<Integer, String> cf =
                            new DefaultKafkaConsumerFactory<Integer, String>(consumerProps);
        ContainerProperties containerProperties = new ContainerProperties(TEMPLATE_TOPIC);
        KafkaMessageListenerContainer<Integer, String> container =
                            new KafkaMessageListenerContainer<>(cf, containerProperties);
        final BlockingQueue<ConsumerRecord<Integer, String>> records = new LinkedBlockingQueue<>();
        container.setupMessageListener(new MessageListener<Integer, String>() {

        	@Override
        	public void onMessage(ConsumerRecord<Integer, String> record) {
                System.out.println(record);
                records.add(record);
            }

        });
        container.setBeanName("templateTests");
        container.start();
        ContainerTestUtils.waitForAssignment(container, embeddedKafka.getPartitionsPerTopic());
        Map<String, Object> senderProps = KafkaTestUtils.senderProps(embeddedKafka.getBrokersAsString());
        ProducerFactory<Integer, String> pf =
                             new DefaultKafkaProducerFactory<Integer, String>(senderProps);
        KafkaTemplate<Integer, String> template = new KafkaTemplate<>(pf);
        template.setDefaultTopic(TEMPLATE_TOPIC);
        template.sendDefault("foo");
        assertThat(records.poll(10, TimeUnit.SECONDS), hasValue("foo"));
        template.sendDefault(0, 2, "bar");
        ConsumerRecord<Integer, String> received = records.poll(10, TimeUnit.SECONDS);
        assertThat(received, hasKey(2));
        assertThat(received, hasPartition(0));
        assertThat(received, hasValue("bar"));
        template.send(TEMPLATE_TOPIC, 0, 2, "baz");
        received = records.poll(10, TimeUnit.SECONDS);
        assertThat(received, hasKey(2));
        assertThat(received, hasPartition(0));
        assertThat(received, hasValue("baz"));
    }

}

The above uses the hamcrest matchers; with AssertJ, the final part looks like this…​

...
        assertThat(records.poll(10, TimeUnit.SECONDS)).has(value("foo"));
        template.sendDefault(0, 2, "bar");
        ConsumerRecord<Integer, String> received = records.poll(10, TimeUnit.SECONDS);
        assertThat(received).has(key(2));
        assertThat(received).has(partition(0));
        assertThat(received).has(value("bar"));
        template.send(TEMPLATE_TOPIC, 0, 2, "baz");
        received = records.poll(10, TimeUnit.SECONDS);
        assertThat(received).has(key(2));
        assertThat(received).has(partition(0));
        assertThat(received).has(value("baz"));
    }
}

This documentation pertains to versions 2.0.0 and above; for documentation for earlier releases, see the 1.3.x README.

Spring Integration Kafka is now based on the Spring for Apache Kafka project. It provides the following components:

These are discussed in the following sections.

The Outbound channel adapter is used to publish messages from a Spring Integration channel to Kafka topics. The channel is defined in the application context and then wired into the application that sends messages to Kafka. Sender applications can publish to Kafka via Spring Integration messages, which are internally converted to Kafka messages by the outbound channel adapter, as follows: the payload of the Spring Integration message will be used to populate the payload of the Kafka message, and (by default) the kafka_messageKey header of the Spring Integration message will be used to populate the key of the Kafka message.

The target topic and partition for publishing the message can be customized through the kafka_topic and kafka_partitionId headers, respectively.

In addition, the <int-kafka:outbound-channel-adapter> provides the ability to extract the key, target topic, and target partition by applying SpEL expressions on the outbound message. To that end, it supports the mutually exclusive pairs of attributes topic/topic-expression, message-key/message-key-expression, and partition-id/partition-id-expression, to allow the specification of topic,message-key and partition-id respectively as static values on the adapter, or to dynamically evaluate their values at runtime against the request message.

Important

The KafkaHeaders interface (provided by spring-kafka) contains constants used for interacting with headers. The messageKey and topic default headers now require a kafka_ prefix. When migrating from an earlier version that used the old headers, you need to specify message-key-expression="headers['messageKey']" and topic-expression="headers['topic']" on the <int-kafka:outbound-channel-adapter>, or simply change the headers upstream to the new headers from KafkaHeaders using a <header-enricher> or MessageBuilder. Or, of course, configure them on the adapter using topic and message-key if you are using constant values.

NOTE : If the adapter is configured with a topic or message key (either with a constant or expression), those are used and the corresponding header is ignored. If you wish the header to override the configuration, you need to configure it in an expression, such as:

topic-expression="headers['topic'] != null ? headers['topic'] : 'myTopic'".

The adapter requires a KafkaTemplate.

Here is an example of how the Kafka outbound channel adapter is configured with XML:

<int-kafka:outbound-channel-adapter id="kafkaOutboundChannelAdapter"
                                    kafka-template="template"
                                    auto-startup="false"
                                    channel="inputToKafka"
                                    topic="foo"
                                    message-key-expression="'bar'"
                                    partition-id-expression="2">
</int-kafka:outbound-channel-adapter>

<bean id="template" class="org.springframework.kafka.core.KafkaTemplate">
    <constructor-arg>
        <bean class="org.springframework.kafka.core.DefaultKafkaProducerFactory">
            <constructor-arg>
                <map>
                    <entry key="bootstrap.servers" value="localhost:9092" />
                    ... <!-- more producer properties -->
                </map>
            </constructor-arg>
        </bean>
    </constructor-arg>
</bean>

As you can see, the adapter requires a KafkaTemplate which, in turn, requires a suitably configured KafkaProducerFactory.

When using Java Configuration:

@Bean
@ServiceActivator(inputChannel = "toKafka")
public MessageHandler handler() throws Exception {
    KafkaProducerMessageHandler<String, String> handler =
            new KafkaProducerMessageHandler<>(kafkaTemplate());
    handler.setTopicExpression(new LiteralExpression("someTopic"));
    handler.setMessageKeyExpression(new LiteralExpression("someKey"));
    return handler;
}

@Bean
public KafkaTemplate<String, String> kafkaTemplate() {
    return new KafkaTemplate<>(producerFactory());
}

@Bean
public ProducerFactory<String, String> producerFactory() {
    Map<String, Object> props = new HashMap<>();
    props.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, this.brokerAddress);
    // set more properties
    return new DefaultKafkaProducerFactory<>(props);
}

The KafkaMessageDrivenChannelAdapter (<int-kafka:message-driven-channel-adapter>) uses a spring-kafka KafkaMessageListenerContainer or ConcurrentListenerContainer.

Starting with spring-integration-kafka version 2.1, the mode attribute is available (record or batch, default record). For record mode, each message payload is converted from a single ConsumerRecord; for mode batch the payload is a list of objects which are converted from all the ConsumerRecord s returned by the consumer poll. As with the batched @KafkaListener, the KafkaHeaders.RECEIVED_MESSAGE_KEY, KafkaHeaders.RECEIVED_PARTITION_ID, KafkaHeaders.RECEIVED_TOPIC and KafkaHeaders.OFFSET headers are also lists with, positions corresponding to the position in the payload.

An example of xml configuration variant is shown here:

<int-kafka:message-driven-channel-adapter
        id="kafkaListener"
        listener-container="container1"
        auto-startup="false"
        phase="100"
        send-timeout="5000"
        mode="record"
        channel="nullChannel"
        error-channel="errorChannel" />

<bean id="container1" class="org.springframework.kafka.listener.KafkaMessageListenerContainer">
    <constructor-arg>
        <bean class="org.springframework.kafka.core.DefaultKafkaConsumerFactory">
            <constructor-arg>
                <map>
                <entry key="bootstrap.servers" value="localhost:9092" />
                ...
                </map>
            </constructor-arg>
        </bean>
    </constructor-arg>
    <constructor-arg>
        <bean class="org.springframework.kafka.listener.config.ContainerProperties">
            <constructor-arg name="topics" value="foo" />
        </bean>
    </constructor-arg>

</bean>

When using Java Configuration:

@Bean
public KafkaMessageDrivenChannelAdapter<String, String>
            adapter(KafkaMessageListenerContainer<String, String> container) {
    KafkaMessageDrivenChannelAdapter<String, String> kafkaMessageDrivenChannelAdapter =
            new KafkaMessageDrivenChannelAdapter<>(container, ListenerMode.record);
    kafkaMessageDrivenChannelAdapter.setOutputChannel(received());
    return kafkaMessageDrivenChannelAdapter;
}

@Bean
public KafkaMessageListenerContainer<String, String> container() throws Exception {
    ContainerProperties properties = new ContainerProperties(this.topic);
    // set more properties
    return new KafkaMessageListenerContainer<>(consumerFactory(), properties);
}

@Bean
public ConsumerFactory<String, String> consumerFactory() {
    Map<String, Object> props = new HashMap<>();
    props.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, this.brokerAddress);
    // set more properties
    return new DefaultKafkaConsumerFactory<>(props);
}

A StringJsonMessageConverter is provided, see Section 4.1.3, “Serialization/Deserialization and Message Conversion” for more information.

When using this converter with a message-driven channel adapter, you can specify the type to which you want the incoming payload to be converted. This is achieved by setting the payload-type attribute (payloadType property) on the adapter.

<int-kafka:message-driven-channel-adapter
        id="kafkaListener"
        listener-container="container1"
        auto-startup="false"
        phase="100"
        send-timeout="5000"
        channel="nullChannel"
        message-converter="messageConverter"
        payload-type="com.example.Foo"
        error-channel="errorChannel" />

<bean id="messageConverter" class="org.springframework.kafka.support.converter.MessagingMessageConverter"/>

@Bean
public KafkaMessageDrivenChannelAdapter<String, String>
            adapter(KafkaMessageListenerContainer<String, String> container) {
    KafkaMessageDrivenChannelAdapter<String, String> kafkaMessageDrivenChannelAdapter =
            new KafkaMessageDrivenChannelAdapter<>(container, ListenerMode.record);
    kafkaMessageDrivenChannelAdapter.setOutputChannel(received());
    kafkaMessageDrivenChannelAdapter.setMessageConverter(converter());
    kafkaMessageDrivenChannelAdapter.setPayloadType(Foo.class);
    return kafkaMessageDrivenChannelAdapter;
}