SwarrotBundle

A bundle to use Swarrot inside your Symfony application.

Installation

The recommended way to install this bundle is through Composer. Just run:

composer require swarrot/swarrot-bundle

Register the bundle in the kernel of your application:

// app/AppKernel.php
public function registerBundles()
{
    $bundles = array(
        // ...
        new Swarrot\SwarrotBundle\SwarrotBundle(),
    );

    return $bundles;
}

Configuration reference

swarrot:
    provider: pecl # pecl or amqp_lib (require php-amqplib/php-amqplib)
    default_connection: rabbitmq
    default_command: swarrot.command.base # Swarrot\SwarrotBundle\Command\SwarrotCommand
    logger: logger # logger or channel logger like monolog.logger.[my_channel]
    connections:
        rabbitmq:
            url: "amqp://%rabbitmq_login%:%rabbitmq_password%@%rabbitmq_host%:%rabbitmq_port%/%rabbitmq_vhost%"
    consumers:
        my_consumer:
            processor: my_consumer.processor.service # Symfony service id implementing Swarrot\Processor\ProcessorInterface
            middleware_stack: # order matters
                 - configurator: swarrot.processor.signal_handler
                   # extras:
                   #     signal_handler_signals:
                   #         - SIGTERM
                   #         - SIGINT
                   #         - SIGQUIT
                 # - configurator: swarrot.processor.insomniac
                 - configurator: swarrot.processor.max_messages
                   # extras:
                   #     max_messages: 100
                 - configurator: swarrot.processor.max_execution_time
                   # extras:
                   #     max_execution_time: 300
                 - configurator: swarrot.processor.memory_limit
                   # extras:
                   #     memory_limit: null
                 - configurator: swarrot.processor.doctrine_connection
                   # extras:
                   #     doctrine_ping: true
                   #     doctrine_close_master: true
                 - configurator: swarrot.processor.doctrine_object_manager
                 - configurator: swarrot.processor.exception_catcher

                 - configurator: swarrot.processor.ack
                   # extras:
                   #     requeue_on_error: false
                 - configurator: swarrot.processor.retry
                   # extras:
                   #     retry_exchange: retry
                   #     retry_attempts: 3
                   #     retry_routing_key_pattern: 'retry_%%attempt%%'

                 # - configurator: swarrot.processor.services_resetter

            extras:
                poll_interval: 500000
    messages_types:
        my_publisher:
            connection: rabbitmq # use the default connection by default
            exchange: my_exchange
            routing_key: my_routing_key

Publishing a message

First step is to retrieve the Swarrot publisher service from your controller.

$messagePublisher = $this->get('swarrot.publisher');

After that, you need to prepare your message with the Message class.

use Swarrot\Broker\Message;

$message = new Message('"My first message with the awesome swarrot lib :)"');

Then you can publish a new message into a predefined configuration (connection, exchange, routing_key, etc.) from your message_types.

$messagePublisher->publish('my_publisher', $message);

When publishing a message, you can override the message_types configuration by passing a third argument:

$messagePublisher->publish('my_publisher', $message, array(
    'exchange'    => 'my_new_echange',
    'connection'  => 'my_second_connection',
    'routing_key' => 'my_new_routing_key'
));

Consuming a message

Swarrot will automatically create one command per consumer defined in your configuration. These command need the queue name to consume as first argument. You can also use a named connection as second argument if you don't want to use the default one.

app/console swarrot:consume:my_consumer queue_name [connection_name]

Your consumer (my_consumer.processor.service) must implements Swarrot\Processor\ProcessorInterface

use Swarrot\Processor\ProcessorInterface;

class MyProcessor implements ProcessorInterface
{
    public function process(Message $message, array $options)
    {
        var_dump($message->getBody()); // "My first message with the awesome swarrot lib :)"
    }
}

Your processor will also be decorated automatically by all processors listed in the middleware_stack section. The order matters.

All these processors are configurable. You can add an extras key on each configurator definition in your config.yml. Take a look at the configuration reference to see available extras for existing Configurators.

You can also use options of the command line:

• --poll-interval [default: 500000]: Change the polling interval when no message found in broker
• --requeue-on-error (-r): Re-queue the message in the same queue if an error occurred.
• --no-catch (-C): Disable the ExceptionCatcher processor (available only if the processor is in the stack)
• --max-execution-time (-t) [default: 300]: Configure the MaxExecutionTime processor (available only if the processor is in the stack)
• --max-messages (-m) [default: 300]: Configure the MaxMessages processor (available only if the processor is in the stack)
• --no-retry (-R): Disable the Retry processor (available only if the processor is in the stack)

Default values will be overriden by your config.yml and usage of options will override default config values.

Run your command with -h to have the full list of options.

Note that you can define one or more aliases for this command using the command_alias configuration:

swarrot:
    consumers:
        my_consumer:
            command_alias: 'my:super:commmand'

Thus allowing you to consume messages using a more appropriate wording:

app/console my:super:command queue_name [connection_name]

Implementing your own Provider

If you want to implement your own provider (like Redis), you first have to implement the Swarrot\SwarrotBundle\Broker\FactoryInterface. Then, you can register it along with the others services and tag it with swarrot.provider_factory.

services:
    app.swarrot.custom_provider_factory:
        class: AppBundle\Provider\CustomFactory
        tags:
            - {name: swarrot.provider_factory}
    app.swarrot.redis_provider_factory:
        class: AppBundle\Provider\RedisFactory
        tags:
            - {name: swarrot.provider_factory, alias: redis}

Now you can tell Swarrot to use it in the config.yml file.

swarrot:
  provider: app.swarrot.custom_provider_factory

or with the alias

swarrot:
  provider: redis

Using a custom processor

If you want to use a custom processor, you need two things. The Processor itself and a ProcessorConfigurator. For the Processor, you can refer to the swarrot/swarrot documentation. For the ConfigurationProcessor, you need to implement the ProcessorConfiguratorInterface and to register it as an abstract service, like this:

services:
  my_own_processor_configurator_service_id:
    abstract: true
    class: MyProject\MyOwnProcessorConfigurator

Once done, just add it to the middleware stack of your consumer:

middleware_stack:
  - configurator: swarrot.processor.signal_handler
  - configurator: my_own_processor_configurator_service_id

As usual, take care of the order of your middleware_stack.

Running your tests without publishing

If you use Swarrot, you may not want to actually publish messages when in test environment for example. You can use the BlackholePublisher to achieve this.

Simply override the swarrot.publisher.class parameter in the DIC with the Swarrot\SwarrotBundle\Broker\BlackholePublisher class, by updating config_test.yml for instance:

parameters:
    swarrot.publisher.class: Swarrot\SwarrotBundle\Broker\BlackholePublisher

Broker configuration

This bundle goal is to deal with message consuming, not to deal with your broker configuration. We don't want to mix the infrastructure logic with the consuming one.

If you're looking for a tool to configure your broker, take a look at odolbeau/rabbit-mq-admin-toolkit.

License

This bundle is released under the MIT License. See the bundled LICENSE file for details.