Produce Avro messages with the schema registry

Set up an Apache Maven project

1. Connect to the client VM using SSH. One way to do this is to run the following command:

   gcloud compute ssh --project=PROJECT_ID \
   --zone=us-central1-f test-instance
   

   For more information about connecting using SSH, see About SSH connections.

2. Install Java and Maven with the command:

   sudo apt-get install maven openjdk-17-jdk
   

3. Set up an Apache Maven project.

   Use the following command to create a package com.google.example in a directory called demo.

   mvn archetype:generate -DartifactId=demo -DgroupId=com.google.example\
      -DarchetypeArtifactId=maven-archetype-quickstart\
      -DarchetypeVersion=1.5 -DinteractiveMode=false
   

Define the schema and its Java implementation

In this example, a message represents a "user" that has a name and an optional ID. The corresponds to an Avro schema with two fields: a required field name of type string and an optional integer id. To use this schema in a Java program, you will also need to generate an Java implementation of an object corresponding to this schema.

1. Change into the project directory with cd demo.

2. Create the folders for storing schema files in your code:

   mkdir -p src/main/avro
   

3. Create the Avro schema definition by pasting the following code into a file called src/main/avro/User.avsc:

   {
     "namespace": "com.google.example",
     "type": "record",
     "name": "User",
     "fields": [
       {"name": "name", "type": "string"},
       {"name": "id",  "type": ["int", "null"]}
     ]
   }
   

4. Configure your Maven project to use an Avro Java code generation plugin by adding the following to the build node of your pom.xml. Note that the pom.xml may have other plugins nodes inside the pluginManagement node. Don't change the pluginManagement node in this step. The plugins node needs to be at the same level level as pluginManagement.

   <plugins>
   <plugin>
     <groupId>org.apache.avro</groupId>
     <artifactId>avro-maven-plugin</artifactId>
     <version>1.11.1</version>
     <executions>
       <execution>
         <phase>generate-sources</phase>
         <goals>
           <goal>schema</goal>
         </goals>
         <configuration>
           <sourceDirectory>${project.basedir}/src/main/avro/</sourceDirectory>
           <outputDirectory>${project.basedir}/src/main/java/</outputDirectory>
         </configuration>
       </execution>
     </executions>
   </plugin>
   </plugins>
   

5. Add Avro as a dependency by adding the following to the end of the project/dependencies node of pom.xml. Note that the pom.xml already has a dependencies node inside dependencyManagement tag. Don't change the dependncyManagement node in this step.

   <dependency>
     <groupId>org.apache.avro</groupId>
     <artifactId>avro</artifactId>
     <version>1.11.1</version>
   </dependency>
   

6. Generate Java sources

     mvn generate-sources
   

7. Run the following command to check that the implementation source file was created. The source is a Java class file that implements constructors, accessors, serializers and de-serializers for User objects. You will use this class in the producer code.

   cat src/main/java/com/google/example/User.java
   

Create a producer client

This section walks through the steps of writing, building, and running a producer client.

Implement the producer

The producer uses KafkaAvroSerializer.java to encode messages and manage their schemas. The serializer automatically connects to the schema registry, registers the schema under a subject, retrieves its ID, and then serializes the message using Avro. You still need to configure the producer and serializer.

1. Create the producer client class by pasting the following code into a new file called src/main/java/com/google/example/UserProducer.java

   
   package com.google.example;
   import java.util.Properties;
   import org.apache.kafka.clients.producer.KafkaProducer;
   import org.apache.kafka.clients.producer.ProducerRecord;
   import org.apache.kafka.clients.producer.RecordMetadata;
   import org.apache.kafka.clients.producer.ProducerConfig;
   import org.apache.kafka.common.serialization.StringSerializer;
   import io.confluent.kafka.serializers.KafkaAvroSerializer;
   
     public class UserProducer {
   
       private static Properties configure() throws Exception {
           Properties p = new Properties();
           p.load(new java.io.FileReader("client.properties"));
           p.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, StringSerializer.class);
           p.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, KafkaAvroSerializer.class);
           return p;
       }
   
       public static void main(String[] args) throws Exception {
           Properties p = configure();
   
           KafkaProducer<String, User> producer = new KafkaProducer<String, User>(p);
           final User u = new User("SchemaEnthusiast", 42);
           final String topicName = "newUsers";
           ProducerRecord<String, User>  message =
             new ProducerRecord<String, User>(topicName, "", u);
           producer.send(message, new SendCallback());
           producer.close();
       }
     }
   

2. Define the callback class in src/main/java/com/google/example/SendCallback.java:

   package com.google.example;
   
   import org.apache.kafka.clients.producer.Callback;
   import org.apache.kafka.clients.producer.RecordMetadata;
   
   class SendCallback implements Callback {
         public void onCompletion(RecordMetadata m, Exception e){
             if (e == null){
               System.out.println("Produced a message successfully.");
             } else {
               System.out.println(e.getMessage());
             }
         }
   }
   

3. To compile this code, you need the org.apache.kafka.clients package and the serializer code. The serializer Maven artifact is distributed through a custom repository. Add the following node to the project node of your pom.xml to configure this repository:

     <repositories>
       <repository>
         <id>confluent</id>
         <name>Confluent</name>
         <url>https://packages.confluent.io/maven/</url>
       </repository>
     </repositories>
   

4. Add the following to the dependencies node in your pom.xml file:

      <dependency>
         <groupId>org.slf4j</groupId>
         <artifactId>slf4j-simple</artifactId>
         <version>1.7.32</version>
       </dependency>
       <dependency>
         <groupId>io.confluent</groupId>
         <artifactId>kafka-avro-serializer</artifactId>
         <version>7.8.1</version>
       </dependency>
       <dependency>
         <groupId>org.apache.kafka</groupId>
         <artifactId>kafka-clients</artifactId>
         <version>3.7.2</version>
       </dependency>
   

5. To make sure all the dependencies are properly resolved, compile the client:

   mvn compile
   

Create a schema registry

To create a schema registry, run the following command:

gcloud beta managed-kafka schema-registries create REGISTRY_ID \
    --location=REGION

Replace the following:

• REGISTRY_ID: a unique identifier for your new schema registry. This forms part of the registry's resource name. The name must start with a letter, contain only letters (a-z, A-Z), numbers (0-9), and underscores (_), and be 63 characters or less.

• REGION: Google Cloud region where the schema registry is going to be created. This location must match the region of the Kafka cluster or clusters using this registry.

The schema definition that you have created is not yet uploaded to the registry. The producer client does this the first time it runs in the following steps.

Configure and run the producer

At this point the producer won't run since it is not fully configured. To configure the producer, provide both the Kafka and schema registry configuration.

1. Create a file called client.properties in the same directory as your pom.xml and add the following content to it:

   bootstrap.servers=bootstrap.CLUSTER_ID.REGION.managedkafka.PROJECT_ID.cloud.goog:9092
   
   security.protocol=SASL_SSL
   sasl.mechanism=OAUTHBEARER
   sasl.login.callback.handler.class=com.google.cloud.hosted.kafka.auth.GcpLoginCallbackHandler
   sasl.jaas.config=org.apache.kafka.common.security.oauthbearer.OAuthBearerLoginModule required;
   
   schema.registry.url=https://managedkafka.googleapis.com/v1/projects/PROJECT_ID/locations/REGION/schemaRegistries/REGISTRY_ID
   bearer.auth.credentials.source=CUSTOM
   bearer.auth.custom.provider.class=com.google.cloud.hosted.kafka.auth.GcpBearerAuthCredentialProvider
   

   Add the Kafka and schema registry authentication handler dependencies to your Maven project by inserting the following to the dependencies node of pom.xml above the kafka-avro-serializer dependency:

     <dependency>
         <groupId>com.google.cloud.hosted.kafka</groupId>
         <artifactId>managed-kafka-auth-login-handler</artifactId>
         <version>1.0.6</version>
         <exclusions>
           <exclusion>
             <groupId>io.confluent</groupId>
             <artifactId>kafka-schema-registry-client</artifactId>
           </exclusion>
       </exclusions>
     </dependency>
   

   If you would like to see the implementation of the custom schema registry authentcation handler authentication handler, look at the GcpBearerAuthCredentialProvider class.

2. Compile and run the producer client:

   mvn compile -q exec:java -Dexec.mainClass=com.google.example.UserProducer
   

   If all goes well, you see the output Produced a message successfully generated by the SendCallback class.

Examine the output

1. Check that the User schema has been registered under a subject name derived from the topic and schema names:

   SR_DOMAIN=https://managedkafka.googleapis.com
   SR_PATH=/v1/projects/PROJECT_ID/locations/REGION
   SR_HOST=$SR_DOMAIN/$SR_PATH/schemaRegistries/REGISTRY_ID/subjects
   
   curl -X GET \
     -H "Content-Type: application/vnd.schemaregistry.v1+json" \
     -H "Authorization: Bearer $(gcloud auth print-access-token)"\
     $SR_HOST
   

   The output of this command should look like this:

   ["newUsers-value"]
   

2. Check that the schema registered in the repository is the same as User:

   curl -X GET \
     -H "Content-Type: application/vnd.schemaregistry.v1+json" \
     -H "Authorization: Bearer $(gcloud auth print-access-token)" \
     $SR_HOST/subjects/newUsers-value/versions/1
   

   The output of the command should look like this:

   {
     "subject": "newUsers-value",
     "version": 1,
     "id": 2,
     "schemaType": "AVRO",
     "schema": "{\"type\":\"record\",\"name\":\"User\",\"namespace\":\"com.google.example\",\"fields\":[{\"name\":\"name\",\"type\":\"string\"},{\"name\":\"id\",\"type\":[\"int\",\"null\"]}]}",
     "references": []
   }