[spiral/queue] Add Retry Strategy Management to the Spiral Queue Component

[butschster]

The Spiral framework's queue component allows developers to push and consume queued tasks, but it currently does not have the ability to manage retry strategies for failed jobs. This can be problematic in cases where jobs fail due to transient errors or external dependencies, as retrying the job can increase the likelihood of a successful outcome.

Proposal

I propose that the Spiral queue component be updated to include retry strategy management using the queue.php configuration file. This would allow developers to define retry strategies for each connection in the queue, as well as a default retry strategy for all connections.

<?php

declare(strict_types=1);

return [
    'default' => env('QUEUE_CONNECTION', 'roadrunner'),
    'connections' => [
        'roadrunner' => [
            'driver' => 'roadrunner',
            'pipeline' => 'amqp',
            'retry_strategy' => new LinearRetryStrategy(maxRetries: 3, delayMilliseconds: 1000),
        ],
    ],

    'retry_strategy' => new MultiplierRetryStrategy(
        maxRetries: 3, 
        delayMilliseconds: 1000, 
        multiplier: 1.5, 
        maxDelayMilliseconds: 3000
    ),
    // ...
];

Benefits

Adding retry strategy management to the Spiral queue component would provide the following benefits:

1. Improved reliability: By automatically retrying failed jobs according to a defined strategy, the framework would increase the likelihood of successful job execution.

2. Flexibility: Developers could define different retry strategies for different connections, allowing them to optimize for specific use cases.

3. Ease of use: The retry strategies could be defined in the queue.php configuration file, which is already used to configure other queue-related settings.