To get started with creating Spring Cloud Stream applications, visit the Spring Initializr and create a new Maven project named "GreetingSource". Select Spring Boot {supported-spring-boot-version} in the dropdown. In the Search for dependencies text box type Stream Rabbit or Stream Kafka depending on what binder you want to use.

Next, create a new class, GreetingSource, in the same package as the GreetingSourceApplication class. Give it the following code:

import org.springframework.cloud.stream.annotation.EnableBinding;
import org.springframework.cloud.stream.messaging.Source;
import org.springframework.integration.annotation.InboundChannelAdapter;

@EnableBinding(Source.class)
public class GreetingSource {

    @InboundChannelAdapter(Source.OUTPUT)
    public String greet() {
        return "hello world " + System.currentTimeMillis();
    }
}

The @EnableBinding annotation is what triggers the creation of Spring Integration infrastructure components. Specifically, it will create a Kafka connection factory, a Kafka outbound channel adapter, and the message channel defined inside the Source interface:

public interface Source {

  String OUTPUT = "output";

  @Output(Source.OUTPUT)
  MessageChannel output();

}

The auto-configuration also creates a default poller, so that the greet() method will be invoked once per second. The standard Spring Integration @InboundChannelAdapter annotation sends a message to the source’s output channel, using the return value as the payload of the message.

To test-drive this setup, run a Kafka message broker. An easy way to do this is to use a Docker image:

# On OS X
$ docker run -p 2181:2181 -p 9092:9092 --env ADVERTISED_HOST=`docker-machine ip \`docker-machine active\`` --env ADVERTISED_PORT=9092 spotify/kafka

# On Linux
$ docker run -p 2181:2181 -p 9092:9092 --env ADVERTISED_HOST=localhost --env ADVERTISED_PORT=9092 spotify/kafka

Build the application:

./mvnw clean package

The consumer application is coded in a similar manner. Go back to Initializr and create another project, named LoggingSink. Then create a new class, LoggingSink, in the same package as the class LoggingSinkApplication and with the following code:

import org.springframework.cloud.stream.annotation.EnableBinding;
import org.springframework.cloud.stream.annotation.StreamListener;
import org.springframework.cloud.stream.messaging.Sink;

@EnableBinding(Sink.class)
public class LoggingSink {

    @StreamListener(Sink.INPUT)
    public void log(String message) {
        System.out.println(message);
    }
}

Build the application:

./mvnw clean package

To connect the GreetingSource application to the LoggingSink application, each application must share the same destination name. Starting up both applications as shown below, you will see the consumer application printing "hello world" and a timestamp to the console:

cd GreetingSource
java -jar target/GreetingSource-0.0.1-SNAPSHOT.jar --spring.cloud.stream.bindings.output.destination=mydest

cd LoggingSink
java -jar target/LoggingSink-0.0.1-SNAPSHOT.jar --server.port=8090 --spring.cloud.stream.bindings.input.destination=mydest

(The different server port prevents collisions of the HTTP port used to service the Spring Boot Actuator endpoints in the two applications.)

The output of the LoggingSink application will look something like the following:

[           main] s.b.c.e.t.TomcatEmbeddedServletContainer : Tomcat started on port(s): 8090 (http)
[           main] com.example.LoggingSinkApplication       : Started LoggingSinkApplication in 6.828 seconds (JVM running for 7.371)
hello world 1458595076731
hello world 1458595077732
hello world 1458595078733
hello world 1458595079734
hello world 1458595080735