Permalink
Fetching contributors…
Cannot retrieve contributors at this time
365 lines (273 sloc) 14.4 KB

OroMessageQueue Bundle

Note: This article is published in the Oro documentation library.

OroMessageQueueBundle incorporates the OroMessageQueue component into OroPlatform and thereby provides message queue processing capabilities for all application components.

Table of Contents

Overview

The bundle integrates OroMessageQueue component. It adds easy to use configuration layer, register services and tie them together, register handy cli commands.

Jobs

The bundle provides an entity and a web gui for the jobs. So the jobs are created in the db and have a web gui where you can monitor jobs status and interrupt jobs.

Usage

First, you have to configure a transport layer and set one to be default. For the config settings

# config/config.yml

oro_message_queue:
    transport:
        default: '%message_queue_transport%'
        '%message_queue_transport%': '%message_queue_transport_config%'
    client: ~

we can configure one of the supported transports via parameters:

DBAL transport

# config/parameters.yml

    message_queue_transport: DBAL
    message_queue_transport_config: ~

DBAL transport options

Once you configured everything you can start producing messages:

<?php

/** @var Oro\Component\MessageQueue\Client\MessageProducer $messageProducer **/
$messageProducer = $container->get('oro_message_queue.message_producer');

$messageProducer->send('aFooTopic', 'Something has happened');

To consume messages you have to first create a message processor:

<?php
use Oro\Component\MessageQueue\Consumption\MessageProcessor;

class FooMessageProcessor implements MessageProcessor, TopicSubscriberInterface
{
    public function process(Message $message, Session $session)
    {
        echo $message->getBody();

        return self::ACK;
        // return self::REJECT; // when the message is broken
        // return self::REQUEUE; // the message is fine but you want to postpone processing
    }

    public static function getSubscribedTopics()
    {
        return ['aFooTopic'];
    }
}

Register it as a container service and subscribe to the topic:

oro_channel.async.change_integration_status_processor:
    class: 'FooMessageProcessor'
    tags:
        - { name: 'oro_message_queue.client.message_processor' }

Now you can start consuming messages:

./bin/console oro:message-queue:consume

Note: Add -vvv to find out what is going while you are consuming messages. There is a lot of valuable debug info there.

Consumer options

  • --message-limit=MESSAGE-LIMIT Consume n messages and exit
  • --time-limit=TIME-LIMIT Consume messages during this time
  • --memory-limit=MEMORY-LIMIT Consume messages until process reaches this memory limit in MB

The --memory-limit option is recommended for the normal consumer usage. If the option is set a consumer checks the used memory amount after each message processing and terminates if it is exceeded. For example if a consumer was run:

./bin/console oro:message-queue:consume --memory-limit=700

then:

  • The consumer processing a message
  • The consumer checks the used memory amount
  • If it exceeds the option value (i.e. 705 MB or 780Mb or 1300 Mb) the consumer terminates (and Supervisord re-runs it)
  • Otherwise it continues message processing.

We recommend to always set this option to the value 2-3 times less than php memory limit. It will help to avoid php memory limit error during message processing.

We recommend to set the --time-limit option to 5-10 minutes if using the DBAL transport to avoid database connection issues

Supervisord

As you read before consumers can normally interrupt the message procession by many reasons. In the all cases above the interrupted consumer should be re-run. So you must keep running oro:message-queue:consume command and to do this best we advise you to delegate this responsibility to Supervisord. With next program configuration supervisord keeps running four simultaneous instances of oro:message-queue:consume command and cares about relaunch if instance has dead by any reason. Pay attention that the program name defined in the [program:oro_message_consumer] must be unique from any other instances deployed on the same supervisord server even if they are for staging purposes only. As an example, set the following programs [program:prod_oro_message_consumer] and [program:dev_oro_message_consumer].

[program:oro_message_consumer]
command=/path/to/bin/console --env=prod --no-debug oro:message-queue:consume
process_name=%(program_name)s_%(process_num)02d
numprocs=4
autostart=true
autorestart=true
startsecs=0
user=apache
redirect_stderr=true

Name prefix for the Message Queue

To use several independent Message Queues on single RabbitMQ instance, configure a name prefix for the Message Queue. For example:

# config/config.yml

oro_message_queue:
    ...
    client:
        prefix: mq_oro_platform_test
        router_destination: queue_name
        default_destination: queue_name

In router_destsination and default_destionation, put the names of the queue specific to your environment. In the prefix option, provide a string that should be prepended to the queue name.

Internals

Structure

You can skip it if you are only going to use the component. The component is split into several layers:

  • Transport - The transport API provides a common way for programs to create, send, receive and read messages. Inspired by Java Message Service
  • Router - An implementation of RecipientList pattern.
  • Consumption - the layer provides tools to simplify consumption of messages. It provides a cli command, a queue consumer, message processor and ways to extend it.
  • Client - provides a high level abstraction. It provides easy to use abstraction for producing and processing messages. It also reduces a need to configure a broker.

Component structure

Flow

The client's message producer sends a message to a router message processor. It takes the message and search for real recipients who is interested in such a message. Then, It sends a copy of a message for all of them. Each target message processor takes its copy of the message and process it.

Message flow

The message itself has headers and body and they change this way while traveling through the system:

Message structure

Custom transport

If you happen to need to implement a custom provider take a look at transport's interfaces. You have to provide an implementation for them

Key Classes

  • MessageProducer - The client's message producer, you will use it all the time to send messages
  • MessageProcessorInterface - Each class which does the job has to implement this interface
  • TopicSubscriberInterface - Kind of EventSubscriberInterface. It allows you to keep a processing code and topics it is subscribed to in one place.
  • MessageConsumeCommand - A command you use to consume messages.
  • QueueConsumer - A class that works inside the command and watch for a new message and once it is get it pass it to a message processor.

Unit and Functional tests

To test that a message was sent in unit and functional tests, you can use MessageQueueExtension trait. There are two implementation of this trait, one for unit tests, another for functional tests:

Also, in case if you need custom logic for manage sent messages, you can use Oro\Bundle\MessageQueueBundle\Test\Unit\MessageQueueAssertTrait or Oro\Bundle\MessageQueueBundle\Test\Functional\MessageQueueAssertTrait traits.

Before you start to use traits in functional tests, you need to register oro_message_queue.test.message_collector service for test environment.

# config/config_test.yml

services:
    oro_message_queue.test.message_collector:
        class: Oro\Bundle\MessageQueueBundle\Test\Functional\MessageCollector
        decorates: oro_message_queue.client.message_producer
        arguments:
            - '@oro_message_queue.test.message_collector.inner'

The following example shows how to test whether a message was sent.

<?php
namespace Acme\Bundle\AcmeBundle\Tests\Functional;

use Oro\Bundle\MessageQueueBundle\Test\Functional\MessageQueueExtension;
use Oro\Bundle\TestFrameworkBundle\Test\WebTestCase;

class SomeTest extends WebTestCase
{
    use MessageQueueExtension;

    public function testSingleMessage()
    {
        // assert that a message was sent to a topic
        self::assertMessageSent('aFooTopic', 'Something has happened');

        // assert that at least one message was sent to a topic
        // can be used if a message is not matter
        self::assertMessageSent('aFooTopic');
    }

    public function testSeveralMessages()
    {
        // assert that exactly given messages were sent to a topic
        self::assertMessagesSent(
            'aFooTopic',
            [
                'Something has happened',
                'Something else has happened',
            ]
        );
        // assert that the exactly given number of messages were sent to a topic
        // can be used if messages are not matter
        self::assertMessagesCount('aFooTopic', 2);
        // also assertCountMessages alias can be used to do the same assertion
        self::assertCountMessages('aFooTopic');
    }

    public function testNoMessages()
    {
        // assert that no any message was sent to a topic
        self::assertMessagesEmpty('aFooTopic');
        // also assertEmptyMessages alias can be used to do the same assertion
        self::assertEmptyMessages('aFooTopic');
    }

    public function testAllMessages()
    {
        // assert that exactly given messages were sent
        // NOTE: use this assertion with caution because it is possible
        // that messages not related to a testing functionality were sent as well
        self::assertAllMessagesSent(
            [
                ['topic' => 'aFooTopic', 'message' => 'Something has happened'],
                ['topic' => 'aFooTopic', 'message' => 'Something else has happened'],
            ]
        );
    }
}

In unit tests you are usually need to pass the message producer to a service you test. To fetch correct instance of message producer in the unit tests use self::getMessageProducer(), e.g.:

<?php
namespace Acme\Bundle\AcmeBundle\Tests\Unit;

use Acme\Bundle\AcmeBundle\SomeClass;
use Oro\Bundle\MessageQueueBundle\Test\Unit\MessageQueueExtension;

class SomeTest extends \PHPUnit\Framework\TestCase
{
    use MessageQueueExtension;

    public function testSingleMessage()
    {
        $instance = new SomeClass(self::getMessageProducer());
        
        $instance->doSomethind();

        self::assertMessageSent('aFooTopic', 'Something has happened');
    }
}

Stale Jobs

It is not possible to create two unique jobs with the same name. That's why if one unique job is not able to finish its work, it can block another job.

To avoid this situation, you can set maximum time for the unique job execution. If the job is still running longer than that, it is possible to create new copy of the unique job (with the same name). The old job is marked as "stale" in this case. See Stale Jobs for more details.

You can configure the time_before_stale parameter in config.yml file, providing time in seconds:

oro_message_queue:
    time_before_stale:
        default: 1800
        jobs:
            bundle_name.processor_name.entity_name.user: 3600
            bundle_name.processor_name.entity_name: 2000
            bundle_name.processor_name: -1

The parser first searches for job by its full name. If the job is not found by the full name, the parser attempts to match the longest part of the job name (reading from the left).

In the example above:

  • bundle_name.processor_name.entity_name.user will be staled after 3600 seconds
  • bundle_name.processor_name.entity_name.organisation will be staled after 2000 seconds
  • bundle_name.processor_name.other_name.some_job will never be staled.
  • bundle_name.other_processor.other_name.some_job will be staled after 1800 seconds
  • processor_name.entity_name.user will be staled after 1800 seconds