Adding a Queue Processor

Logidav uses a SQL-backed queue system (with optional RabbitMQ) for asynchronous task processing. This guide covers creating a new queue processor that handles queued work items.

How the Queue System Works

The central command meduse:queue:processor --action=runQueues polls for pending queue entries and dispatches each one to the appropriate processor based on the model field. The run_category field controls scheduling windows so that heavy tasks run during off-peak hours.

1. Implement QueueProcessorInterface

Create your processor class in the appropriate bundle's Services/Queue/ directory:

<?php
// src/AppBundle/Services/Queue/OrderExportProcessor.php

namespace AppBundle\Services\Queue;

use MeduseBundle\Services\Queue\QueueProcessorInterface;
use Doctrine\ORM\EntityManagerInterface;
use Psr\Log\LoggerInterface;

class OrderExportProcessor implements QueueProcessorInterface
{
    private $em;
    private $logger;

    public function __construct(EntityManagerInterface $em, LoggerInterface $logger)
    {
        $this->em = $em;
        $this->logger = $logger;
    }

    /**
     * Process a batch of queue entries.
     *
     * @param array $queues Array of Queue entities to process
     * @return array Results keyed by queue ID
     */
    public function execute(array $queues): array
    {
        $results = [];

        foreach ($queues as $queue) {
            try {
                $payload = json_decode($queue->getPayload(), true);

                $this->logger->info('Processing order export', [
                    'queue_id' => $queue->getId(),
                    'order_id' => $payload['order_id'] ?? null,
                ]);

                // Your processing logic here
                $this->exportOrder($payload);

                $results[$queue->getId()] = [
                    'status' => 'success',
                    'message' => 'Order exported successfully',
                ];
            } catch (\Exception $e) {
                $this->logger->error('Order export failed', [
                    'queue_id' => $queue->getId(),
                    'error' => $e->getMessage(),
                ]);

                $results[$queue->getId()] = [
                    'status' => 'error',
                    'message' => $e->getMessage(),
                ];
            }
        }

        return $results;
    }

    private function exportOrder(array $payload)
    {
        // Business logic for exporting the order
    }
}

2. Register the Processor

Add the service definition in your bundle's services.yml. The service ID is important -- it becomes the model field stored in the Queue entity:

services:
    app.queue.order_export:
        class: AppBundle\Services\Queue\OrderExportProcessor
        arguments:
            - '@doctrine.orm.entity_manager'
            - '@logger'

When creating queue entries, set the model field to match this service ID:

$queue = new Queue();
$queue->setModel('app.queue.order_export');
$queue->setPayload(json_encode(['order_id' => $orderId]));
$queue->setRunCategory('standard');
$this->em->persist($queue);
$this->em->flush();

3. Queue Categories

The run_category field controls when and how queues are processed:

Category	Description	Typical Schedule
standard	Normal priority, runs frequently	Every 1-5 minutes
heavy	Resource-intensive tasks	Off-peak hours
realtime	Urgent tasks needing immediate processing	Every minute

The queue processor command filters by category:

# Process standard queues
php bin/console meduse:queue:processor --action=runQueues --category=standard

# Process heavy queues (typically scheduled during off-peak)
php bin/console meduse:queue:processor --action=runQueues --category=heavy

4. Error Handling

Robust error handling is critical for queue processors. The system tracks retries automatically:

• Catch exceptions at the individual queue entry level so one failure does not block others.
• Log with context -- include the queue ID, relevant entity IDs, and error details.
• Update queue status -- return error status in results so the framework marks the entry correctly.
• Retry tracking -- the number_of_trials field on the Queue entity increments on each attempt. Check it to implement maximum retry logic:

if ($queue->getNumberOfTrials() >= 3) {
    $this->logger->warning('Queue entry exceeded max retries, marking as failed', [
        'queue_id' => $queue->getId(),
    ]);
    $results[$queue->getId()] = [
        'status' => 'failed',
        'message' => 'Max retries exceeded',
    ];
    continue;
}

warning

Always wrap individual queue entry processing in try/catch. An uncaught exception in one entry must not prevent processing of the remaining entries in the batch.

5. Testing

Unit Test

Test your processor logic in isolation by mocking dependencies:

<?php
// tests/AppBundle/Services/Queue/OrderExportProcessorTest.php

namespace Tests\AppBundle\Services\Queue;

use AppBundle\Services\Queue\OrderExportProcessor;
use PHPUnit\Framework\TestCase;

class OrderExportProcessorTest extends TestCase
{
    public function testExecuteProcessesQueuesSuccessfully()
    {
        $em = $this->createMock(\Doctrine\ORM\EntityManagerInterface::class);
        $logger = $this->createMock(\Psr\Log\LoggerInterface::class);

        $processor = new OrderExportProcessor($em, $logger);

        $queue = $this->createMock(\MeduseBundle\Entity\Queue::class);
        $queue->method('getId')->willReturn(1);
        $queue->method('getPayload')->willReturn(json_encode(['order_id' => 42]));

        $results = $processor->execute([$queue]);

        $this->assertArrayHasKey(1, $results);
        $this->assertEquals('success', $results[1]['status']);
    }
}

Integration Test

Create test queue entries in the database and verify end-to-end processing:

# Insert a test queue entry, then run the processor
php bin/console meduse:queue:processor --action=runQueues -e=dev

Checklist

• Processor class implements QueueProcessorInterface
• Service registered in services.yml
• Queue entries use the correct model (service ID) and run_category
• Error handling wraps individual entries in try/catch
• Retry logic respects number_of_trials
• Unit tests cover success and failure paths
• Crontab entry added if a new category/schedule is needed

See Also

• Architecture: Queue Model for queue internals
• Adding a Cronjob for scheduling the queue processor command
• Adding an Integration if the processor calls external APIs