mateusjunges/laravel-kafka

A kafka driver for laravel

v1.6.1 2022-01-20 20:50 UTC

README

docs/laravel-kafka.png

Latest Version On Packagist Total Downloads MIT Licensed Continuous Integration Check & fix styling PHP Version Require

Do you use Kafka in your laravel projects? All packages I've seen until today, including some built by myself, does not provide a nice syntax usage syntax or, if it does, the test process with these packages are very painful.

This package provides a nice way of producing and consuming kafka messages in your Laravel projects.

Follow these docs to install this package and start using kafka in your laravel projects.

Installation

To install this package, you must have installed PHP RdKafka extension. You can follow the steps here to install rdkafka in your system.

With RdKafka installed, require this package with composer:

composer require mateusjunges/laravel-kafka

You can publish the package configuration using:

php artisan vendor:publish --tag=laravel-kafka-config

Now you are good to go!

Usage

After installing the package, you can start producing and consuming Kafka messages.

Producing Kafka Messages

To publish your messages to Kafka, you can use the publishOn method, of Junges\Kafka\Facades\Kafka class:

use Junges\Kafka\Facades\Kafka;

Kafka::publishOn('topic');

You can also specify the broker where you want to publish the message:

use Junges\Kafka\Facades\Kafka;

Kafka::publishOn('topic', 'broker');

This method returns a Junges\Kafka\Producers\ProducerBuilder::class instance, and you can configure your message.

The ProducerBuilder class contains a few methods to configure your kafka producer. The following lines describes these methods.

ProducerBuilder configuration methods

The withConfigOption method sets a \RdKafka\Conf::class option. You can check all available options here. This methods set one config per call, and you can use withConfigOptions passing an array of config name and config value as argument. Here's an example:

use Junges\Kafka\Facades\Kafka;

Kafka::publishOn('topic')
    ->withConfigOption('property-name', 'property-value')
    ->withConfigOptions([
        'property-name' => 'property-value'
    ]);

While you are developing your application, you can enable debug with the withDebugEnabled method. To disable debug mode, you can use ->withDebugEnabled(false), or withDebugDisabled methods.

use Junges\Kafka\Facades\Kafka;

Kafka::publishOn('topic')
    ->withConfigOption('property-name', 'property-value')
    ->withConfigOptions([
        'property-name' => 'property-value'
    ])
    ->withDebugEnabled() // To enable debug mode
    ->withDebugDisabled() // To disable debug mode
    ->withDebugEnabled(false) // Also to disable debug mode

Using custom serializers

To use custom serializers, you must use the usingSerializer method:

$producer = \Junges\Kafka\Facades\Kafka::publishOn('topic')->usingSerializer(new MyCustomSerializer());

Using AVRO serializer

To use the AVRO serializer, add the AVRO serializer:

use FlixTech\AvroSerializer\Objects\RecordSerializer;
use FlixTech\SchemaRegistryApi\Registry\CachedRegistry;
use FlixTech\SchemaRegistryApi\Registry\BlockingRegistry;
use FlixTech\SchemaRegistryApi\Registry\PromisingRegistry;
use FlixTech\SchemaRegistryApi\Registry\Cache\AvroObjectCacheAdapter;
use GuzzleHttp\Client;

$cachedRegistry = new CachedRegistry(
    new BlockingRegistry(
        new PromisingRegistry(
            new Client(['base_uri' => 'kafka-schema-registry:9081'])
        )
    ),
    new AvroObjectCacheAdapter()
);

$registry = new AvroSchemaRegistry($cachedRegistry);
$recordSerializer = new RecordSerializer($cachedRegistry);

//if no version is defined, latest version will be used
//if no schema definition is defined, the appropriate version will be fetched form the registry
$registry->addBodySchemaMappingForTopic(
    'test-topic',
    new \Junges\Kafka\Message\KafkaAvroSchema('bodySchemaName' /*, int $version, AvroSchema $definition */)
);
$registry->addKeySchemaMappingForTopic(
    'test-topic',
    new \Junges\Kafka\Message\KafkaAvroSchema('keySchemaName' /*, int $version, AvroSchema $definition */)
);

$serializer = new \Junges\Kafka\Message\Serializers\AvroSerializer($registry, $recordSerializer /*, AvroEncoderInterface::ENCODE_BODY */);

$producer = \Junges\Kafka\Facades\Kafka::publishOn('topic')->usingSerializer($serializer);

Configuring the Kafka message payload

In kafka, you can configure your payload with a message, message headers and message key. All these configurations are available within ProducerBuilder class.

Configuring message headers

To configure the message headers, use the withHeaders method:

use Junges\Kafka\Facades\Kafka;

Kafka::publishOn('topic')
    ->withHeaders([
        'header-key' => 'header-value'
    ])

Configure the message body

You can configure the message with the withMessage or withBodyKey methods.

The withMessage sets the entire message, and it accepts a Junges\Kafka\Message\Message::class instance as argument.

This is how you should use it:

use Junges\Kafka\Facades\Kafka;
use Junges\Kafka\Message\Message;

$message = new Message(
    headers: ['header-key' => 'header-value'],
    body: ['key' => 'value'],
    key: 'kafka key here'  
)

Kafka::publishOn('topic')->withMessage($message);

The withBodyKey method sets only a key in your message.

use Junges\Kafka\Facades\Kafka;

Kafka::publishOn('topic')->withBodyKey('key', 'value');

Using Kafka Keys

In Kafka, keys are used to determine the partition within a log to which a message get's appended to. If you want to use a key in your message, you should use the withKafkaKey method:

use Junges\Kafka\Facades\Kafka;

Kafka::publishOn('topic')->withKafkaKey('your-kafka-key');

Sending the message to Kafka

After configuring all your message options, you must use the send method, to send the message to kafka.

use Junges\Kafka\Facades\Kafka;

/** @var \Junges\Kafka\Producers\ProducerBuilder $producer */
$producer = Kafka::publishOn('topic')
    ->withConfigOptions(['key' => 'value'])
    ->withKafkaKey('kafka-key')
    ->withHeaders(['header-key' => 'header-value']);

$producer->send();

Consuming Kafka Messages

If your application needs to read messages from a Kafka topic, you must create a consumer object, subscribe to the appropriate topic and start receiving messages.

To create a consumer using this package, you can use the createConsumer method, on Kafka facade:

use Junges\Kafka\Facades\Kafka;

$consumer = Kafka::createConsumer();

This method also allows you to specify the topics it should consume, the broker and the consumer group id:

use Junges\Kafka\Facades\Kafka;

$consumer = Kafka::createConsumer(['topic-1', 'topic-2'], 'group-id', 'broker');

This method returns a Junges\Kafka\Consumers\ConsumerBuilder::class instance, and you can use it to configure your consumer.

Subscribing to a topic

With a consumer created, you can subscribe to a kafka topic using the subscribe method:

use Junges\Kafka\Facades\Kafka;

$consumer = Kafka::createConsumer()->subscribe('topic');

Of course, you can subscribe to more than one topic at once, either using an array of topics or specifying one by one:

use Junges\Kafka\Facades\Kafka;

$consumer = Kafka::createConsumer()->subscribe('topic-1', 'topic-2', 'topic-n');

// Or, using array:
$consumer = Kafka::createConsumer()->subscribe([
    'topic-1',
    'topic-2',
    'topic-n'
]);

Configuring consumer groups

Kafka consumers belonging to the same consumer group share a group id. THe consumers in a group divides the topic partitions as fairly amongst themselves as possible by establishing that each partition is only consumed by a single consumer from the group.

To attach your consumer to a consumer group, you can use the method withConsumerGroupId to specify the consumer group id:

use Junges\Kafka\Facades\Kafka;

$consumer = Kafka::createConsumer()->withConsumerGroupId('foo');

Configuring message handlers

Now that you have created your kafka consumer, you must create a handler for the messages this consumer receives. By default, a consumer is any callable. You can use an invokable class or a simple callback. Use the withHandler method to specify your handler:

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer();

// Using callback:
$consumer->withHandler(function(\Junges\Kafka\Contracts\KafkaConsumerMessage $message) {
    // Handle your message here
});

Or, using a invokable class:

class Handler
{
    public function __invoke(\Junges\Kafka\Contracts\KafkaConsumerMessage $message){
        // Handle your message here
    }
}

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->withHandler(new Handler)

The KafkaConsumerMessage contract gives you some handy methods to get the message properties:

  • getKey(): Returns the Kafka Message Key
  • getTopicName(): Returns the topic where the message was published
  • getPartition(): Returns the kafka partition where the message was published
  • getHeaders(): Returns the kafka message headers
  • getBody(): Returns the body of the message
  • getOffset(): Returns the offset where the message was published

Configuring max messages to be consumed

If you want to consume a limited amount of messages, you can use the withMaxMessages method to set the max number of messages to be consumed by a kafka consumer:

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->withMaxMessages(2);

Stopping consumer using pcntl

Stopping consumers is very useful if you want to ensure you don't kill a process halfway through processing a consumed message.

To stop the consumer gracefully call the stopConsume method on a consumer instance.

This is particularly useful when using signal handlers. NOTE You will require the Process Control Extension to be installed to utilise the pcntl methods.

function gracefulShutdown(Consumer $consumer) {
    $consumer->stopConsume(function() {
        echo 'Stopped consuming';
        exit(0);
    });
}

$consumer = Kafka::createConsumer(['topic'])
    ->withConsumerGroupId('group')
    ->withHandler(new Handler)
    ->build();
    
pcntl_signal(SIGINT, fn() => gracefulShutdown($consumer));

$consumer->consume();

Configuring a dead letter queue

In kafka, a Dead Letter Queue (or DLQ), is a simple kafka topic in the kafka cluster which acts as the destination for messages that were not able to make it to the desired destination due to some error.

To create a dlq in this package, you can use the withDlq method. If you don't specify the DLQ topic name, it will be created based on the topic you are consuming, adding the -dlq suffix to the topic name.

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->withDlq();

//Or, specifying the dlq topic name:
$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->withDlq('your-dlq-topic-name')

Using SASL

SASL allows your producers and your consumers to authenticate to your Kafka cluster, which verifies their identity. It's also a secure way to enable your clients to endorse an identity. To provide SASL configuration, you can use the withSasl method, passing a Junges\Kafka\Config\Sasl instance as the argument:

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()
    ->withSasl(new \Junges\Kafka\Config\Sasl(
        password: 'password',
        username: 'username'
        mechanisms: 'authentication mechanism'
    ));

You can also set the security protocol used with sasl. It's optional and by default SASL_PLAINTEXT is used, but you can set it to SASL_SSL:

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()
    ->withSasl(new \Junges\Kafka\Config\Sasl(
        password: 'password',
        username: 'username'
        mechanisms: 'authentication mechanism',
        securityProtocol: 'SASL_SSL',
    ));

Using middlewares

Middlewares provides a convenient way to filter and inspecting your Kafka messages. To write a middleware in this package, you can use the withMiddleware method. The middleware is a callable in which the first argument is the message itself and the second one is the next handler. The middlewares get executed in the order they are defined,

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()
    ->withMiddleware(function($message, callable $next) {
        // Perform some work here
        return $next($message);
    });

Using custom deserializers

To set the deserializer you want to use, use the usingDeserializer method:

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->usingDeserializer(new MyCustomDeserializer());

NOTE: The deserializer class must use the same algorithm as the serializer used to produce this message.

Using AVRO deserializer

To use the AVRO deserializer on your consumer, add the Avro deserializer:

use FlixTech\AvroSerializer\Objects\RecordSerializer;
use FlixTech\SchemaRegistryApi\Registry\CachedRegistry;
use FlixTech\SchemaRegistryApi\Registry\BlockingRegistry;
use FlixTech\SchemaRegistryApi\Registry\PromisingRegistry;
use FlixTech\SchemaRegistryApi\Registry\Cache\AvroObjectCacheAdapter;
use GuzzleHttp\Client;


$cachedRegistry = new CachedRegistry(
    new BlockingRegistry(
        new PromisingRegistry(
            new Client(['base_uri' => 'kafka-schema-registry:9081'])
        )
    ),
    new AvroObjectCacheAdapter()
);

$registry = new \Junges\Kafka\Message\Registry\AvroSchemaRegistry($cachedRegistry);
$recordSerializer = new RecordSerializer($cachedRegistry);

//if no version is defined, latest version will be used
//if no schema definition is defined, the appropriate version will be fetched form the registry
$registry->addBodySchemaMappingForTopic(
    'test-topic',
    new \Junges\Kafka\Message\KafkaAvroSchema('bodySchema' , 9 /* , AvroSchema $definition */)
);
$registry->addKeySchemaMappingForTopic(
    'test-topic',
    new \Junges\Kafka\Message\KafkaAvroSchema('keySchema' , 9 /* , AvroSchema $definition */)
);

// if you are only decoding key or value, you can pass that mode as additional third argument
// per default both key and body will get decoded
$deserializer = new \Junges\Kafka\Message\Deserializers\AvroDeserializer($registry, $recordSerializer /*, AvroDecoderInterface::DECODE_BODY */);

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->usingDeserializer($deserializer);

Using auto commit

The auto-commit check is called in every poll and it checks that the time elapsed is greater than the configured time. To enable auto commit, use the withAutoCommit method:

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->withAutoCommit();

Using custom committers

By default the committers provided by the DefaultCommitterFactory are provided.

To set a custom committer on your consumer, add the committer via a factory that implements the CommitterFactory interface:

use \RdKafka\KafkaConsumer;
use \RdKafka\Message;
use \Junges\Kafka\Commit\Contracts\Committer;
use \Junges\Kafka\Commit\Contracts\CommitterFactory;
use \Junges\Kafka\Config\Config;

class MyCommitter implements Committer
{
    public function commitMessage(Message $message, bool $success) : void {
        // ...
    }
    
    public function commitDlq(Message $message) : void {
        // ...
    }  
}

class MyCommitterFactory implements CommitterFactory
{
    public function make(KafkaConsumer $kafkaConsumer, Config $config) : Committer {
        // ...
    }
}

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()
    ->usingCommitterFactory(new MyCommitterFactory())
    ->build();

Setting Kafka configuration options

To set configuration options, you can use two methods: withOptions, passing an array of option and option value or, using the `withOption method and passing two arguments, the option name and the option value.

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()
    ->withOptions([
        'option-name' => 'option-value'
    ]);
// Or:
$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()
    ->withOption('option-name', 'option-value');

Building the consumer

When you have finished configuring your consumer, you must call the build method, which returns a Junges\Kafka\Consumers\Consumer instance.

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()
    // Configure your consumer here
    ->build();

Consuming the kafka messages

After building the consumer, you must call the consume method to consume the messages:

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->build();

$consumer->consume();

Using the built in consumer command

This library provides you a built in consumer command, which you can use to consume messages.

To use this command, you must create a Consumer class, which extends Junges\Kafka\Contracts\Consumer.

Then, just use the following command:

php artisan kafka:consume --consumer=\\App\\Path\\To\\Your\\Consumer --topics=topic-to-consume

Using custom serializers/deserializers

Serialization is the process of converting messages to bytes. Deserialization is the inverse process - converting a stream of bytes into and object. In a nutshell, it transforms the content into readable and interpretable information. Basically, in order to prepare the message for transmission from the producer we use serializers. This package supports three serializers out of the box:

  • NullSerializer / NullDeserializer
  • JsonSerializer / JsonDeserializer
  • AvroSerializer / JsonDeserializer

Changing default serializers and deserializers

The default Serializer is resolved using the MessageSerializer and MessageDeserializer contracts. Out of the box, the Json serializers are used.

To set the default serializer, you can bind the MessageSerializer and MessageDeserializer contracts to any class which implements this interfaces.

Open your AppServiceProvider class and add this lines to the register method:

$this->app->bind(\Junges\Kafka\Contracts\MessageSerializer::class, function () {
   return new MyCustomSerializer();
});

$this->app->bind(\Junges\Kafka\Contracts\MessageDeserializer::class, function() {
    return new MyCustomDeserializer();
});

Creating a custom serializer

To create a custom serializer, you need to create a class that implements the \Junges\Kafka\Contracts\MessageSerializer contract. This interface force you to declare the serialize method.

Creating a custom deserializer

To create a custom deserializer, you need to create a class that implements the \Junges\Kafka\Contracts\MessageDeserializer contract. This interface force you to declare the deserialize method.

Replacing serializers and deserializers on the fly

Serializers and deserializers need to be set both on the Producer and the Consumer classes. To set the producer serializer, you must use the usingSerializer method, available on the ProducerBuilder class. To set the consumer deserializer, you must use the usingDeserializer method, available on the ConsumerBuilder class.

$producer = \Junges\Kafka\Facades\Kafka::publishOn('topic')->usingSerializer(new MyCustomSerializer());

$consumer = \Junges\Kafka\Facades\Kafka::createConsumer()->usingDeserializer(new MyCustomDeserializer());

Using Kafka::fake()

When testing your application, you may wish to "mock" certain aspects of the app, so they are not actually executed during a given test. This package provides convenient helpers for mocking the kafka producer out of the box. These helpers primarily provide a convenience layer over Mockery so you don't have to manually make complicated Mockery method calls.

The Kafka facade also provides methods to perform assertions over published messages, such as assertPublished, assertPublishedOn and assertNothingPublished.

assertPublished method

use Junges\Kafka\Facades\Kafka;
use PHPUnit\Framework\TestCase;

class MyTest extends TestCase
{
     public function testMyAwesomeApp()
     {
         Kafka::fake();
         
         $producer = Kafka::publishOn('topic')
             ->withHeaders(['key' => 'value'])
             ->withBodyKey('foo', 'bar');
             
         $producer->send();
             
         Kafka::assertPublished($producer->getMessage());       
     }
}

You can also use assertPublished without passing the message argument:

use Junges\Kafka\Facades\Kafka;
use PHPUnit\Framework\TestCase;

class MyTest extends TestCase
{
     public function testMyAwesomeApp()
     {
         Kafka::fake();
         
         Kafka::publishOn('topic')
             ->withHeaders(['key' => 'value'])
             ->withBodyKey('foo', 'bar');
             
             
         Kafka::assertPublished();       
     }
}

assertPublishedOn method

If you want to assert that a message was published in a specific kafka topic, you can use the assertPublishedOn method:

use PHPUnit\Framework\TestCase;
use Junges\Kafka\Facades\Kafka;

class MyTest extends TestCase
{
    public function testWithSpecificTopic()
    {
        Kafka::fake();
        
        $producer = Kafka::publishOn('some-kafka-topic')
            ->withHeaders(['key' => 'value'])
            ->withBodyKey('key', 'value');
            
        $producer->send();
        
        Kafka::assertPublishedOn('some-kafka-topic', $producer->getMessage());
        
        // Or:
        Kafka::assertPublishedOn('some-kafka-topic');
        
    }
}

You can also use a callback function to perform assertions within the message using a callback in which the argument is the published message itself.

use PHPUnit\Framework\TestCase;
use Junges\Kafka\Facades\Kafka;
use Junges\Kafka\Message\Message;

class MyTest extends TestCase
{
    public function testWithSpecificTopic()
    {
        Kafka::fake();
        
        $producer = Kafka::publishOn('some-kafka-topic')
            ->withHeaders(['key' => 'value'])
            ->withBodyKey('key', 'value');
            
        $producer->send();
        
        Kafka::assertPublishedOn('some-kafka-topic', $producer->getMessage(), function(Message $message) {
            return $message->getHeaders()['key'] === 'value';
        });
        
        // Or:
        Kafka::assertPublishedOn('some-kafka-topic', null, function(Message $message) {
            return $message->getHeaders()['key'] === 'value';
        });
    }
} 

assertNothingPublished method

You can also assert that nothing was published at all, using the assertNothingPublished:

use PHPUnit\Framework\TestCase;
use Junges\Kafka\Facades\Kafka;
use Junges\Kafka\Message\Message;

class MyTest extends TestCase
{
    public function testWithSpecificTopic()
    {
        Kafka::fake();
        
        if (false) {
            $producer = Kafka::publishOn('some-kafka-topic')
                ->withHeaders(['key' => 'value'])
                ->withBodyKey('key', 'value');
                
            $producer->send();
        }
        
        Kafka::assertNothingPublished();
    }
} 

assertPublishedTimes method

Sometimes, you need to assert that Kafka has published a given number of messages. For that, you can use the assertPublishedTimes method:

use PHPUnit\Framework\TestCase;
use Junges\Kafka\Facades\Kafka;
use Junges\Kafka\Message\Message;

class MyTest extends TestCase
{
    public function testWithSpecificTopic()
    {
        Kafka::fake();

        Kafka::publishOn('some-kafka-topic')
            ->withHeaders(['key' => 'value'])
            ->withBodyKey('key', 'value');

        Kafka::publishOn('some-kafka-topic')
            ->withHeaders(['key' => 'value'])
            ->withBodyKey('key', 'value');

        Kafka::assertPublishedTimes(2);
    }
} 

assertPublishedOnTimes method

To assert that messages were published on a given topic a given number of times, you can use the assertPublishedOnTimes method:

use PHPUnit\Framework\TestCase;
use Junges\Kafka\Facades\Kafka;
use Junges\Kafka\Message\Message;

class MyTest extends TestCase
{
    public function testWithSpecificTopic()
    {
        Kafka::fake();

        Kafka::publishOn('some-kafka-topic')
            ->withHeaders(['key' => 'value'])
            ->withBodyKey('key', 'value');

        Kafka::publishOn('some-kafka-topic')
            ->withHeaders(['key' => 'value'])
            ->withBodyKey('key', 'value');

        Kafka::assertPublishedOnTimes('some-kafka-topic', 2);
    }
} 

Testing

Run composer test to test this package.

Contributing

Thank you for considering contributing for the Laravel Kafka package! The contribution guide can be found here.

Credits

License

The Laravel Kafka package is open-sourced software licenced under the MIT License. Please see the License File for more information.