Bref Messenger failure strategies

So you have fallen in love with Bref and you really want to use Symfony's excellent Messenger component. You've probably also installed the Bref Symfony Messenger bundle that allows you to publish messages on SQS and SNS etc. But you are missing something... You want to be able to use Symfony Messenger retry strategies, right?

This is the package for you!

Install

composer require happyr/bref-messenger-failure-strategies

Now you have a class called Happyr\BrefMessenger\SymfonyBusDriver that implements Bref\Symfony\Messenger\Service\BusDriver. Feel free to configure your consumers with this new class.

Example

On each consumer you can choose to let Symfony handle failures as described in the documentation.

# config/packages/messenger.yaml

framework:
    messenger:
        failure_transport: failed
        transports:
            failed: 'doctrine://default?queue_name=failed'
            workqueue:
              dsn: 'https://sqs.us-east-1.amazonaws.com/123456789/my-queue'
              retry_strategy:
                  max_retries: 3
                  # milliseconds delay
                  delay: 1000
                  multiplier: 2
                  max_delay: 60

services:
    Happyr\BrefMessenger\ExceptionLogger:
        autowire: true
        autoconfigure: true

    Happyr\BrefMessenger\SymfonyBusDriver:
        autowire: true

    Bref\Symfony\Messenger\Service\Sqs\SqsConsumer:
        arguments:
            - '@Happyr\BrefMessenger\SymfonyBusDriver'
            - '@messenger.routable_message_bus'
            - '@Symfony\Component\Messenger\Transport\Serialization\SerializerInterface'
            - 'my_sqs' # Same as transport name

# ...

The delay is only supported on SQS "normal queue". If you are using SNS or SQS FIFO you should use the failure queue directly.

# config/packages/messenger.yaml

framework:
    messenger:
        failure_transport: failed
        transports:
            failed: 'doctrine://default?queue_name=failed'
            workqueue:
              dsn: 'sns://arn:aws:sns:us-east-1:1234567890:foobar'
              retry_strategy:
                  max_retries: 0
services:
    # ...

Make sure you re-run the failure queue time to time. The following config will run a script for 5 seconds every 30 minutes. It will run for 5 seconds even though no messages has failed.

# serverless.yml

functions:
    website:
        # ...
    consumer:
        # ...

    console:
        handler: bin/console
        Timeout: 120 # in seconds
        layers:
            - ${bref:layer.php-74}
            - ${bref:layer.console}
        events:
            - schedule:
                  rate: rate(30 minutes)
                  input:
                      cli: messenger:consume failed --time-limit=5 --limit=50