phobetor/rabbitmq-supervisor-bundle

Symfony bundle to automatically create and update supervisor configurations for RabbitMQ consumer daemons

RabbitMQ supervisor bundle

, _(*1)

Symfony bundle to automatically create and update supervisor configurations for php-amqplib/rabbitmq-bundle (and its predecessor oldsound/rabbitmq-bundle) RabbitMQ consumer daemons., _(*2)

In a nutshell | tl;dr

If you use php-amqplib/rabbitmq-bundle to handle the communication with RabbitMQ, just install supervisor, add this bundle and run, _(*3)

$ app/console rabbitmq-supervisor:rebuild

to get a running supervisord instance that automatically manages all your consumer daemons. When your worker configuration or your code changes, run the command again and all the daemons will be updated., _(*4)

Installation

Install supervisor. e. g. on debian based distributions via apt-get:, _(*5)

# apt-get install supervisor

Add bundle via composer, _(*6)

$ php composer require phobetor/rabbitmq-supervisor-bundle

This will install the bundle to your project’s vendor directory., _(*7)

If your are not using Symfony Flex, also add the bundle to your project’s AppKernel:, _(*8)

// app/AppKernel.php

public function registerBundles()
{
    $bundles = [
        // […]
        new Phobetor\RabbitMqSupervisorBundle\RabbitMqSupervisorBundle(),
    ];
}

Symfony 5:, _(*9)

// config/bundles.php
return [
    ...
    Phobetor\RabbitMqSupervisorBundle\RabbitMqSupervisorBundle::class => ["all" => true],
    ...
];

Zero Configuration

RabbitMQ supervisor bundle works out of the box with a predefined configuration. If you leave it this way you will end up with this directory structure:, _(*10)

supervisor/
└── dev
    ├── logs
    │   ├── stderr.log
    │   └── stdout.log
    ├── supervisord.conf
    ├── supervisord.log
    ├── supervisor.pid
    ├── supervisor.sock
    └── worker
        ├── queue1.conf
        ├── queue2.conf
        ├── queue3.conf
        └── queue4.conf

If you want to use the standard structure inside of the var/ directory, change the configuration like this:, _(*11)

rabbit_mq_supervisor:
    paths:
        workspace_directory:            "%kernel.project_dir%/var/supervisor/%kernel.environment%/"
        configuration_file:             "%kernel.project_dir%/var/supervisor/%kernel.environment%/supervisord.conf"
        pid_file:                       "%kernel.project_dir%/var/supervisor/%kernel.environment%/supervisor.pid"
        sock_file:                      "%kernel.project_dir%/var/supervisor/%kernel.environment%/supervisor.sock"
        log_file:                       "%kernel.project_dir%/var/supervisor/%kernel.environment%/supervisord.log"
        worker_configuration_directory: "%kernel.project_dir%/var/supervisor/%kernel.environment%/worker/"
        worker_output_log_file:         "%kernel.project_dir%/var/supervisor/%kernel.environment%/logs/stdout.log"
        worker_error_log_file:          "%kernel.project_dir%/var/supervisor/%kernel.environment%/logs/stderr.log"

Advanced configuration

To see all configuration options run, _(*12)

$ console config:dump-reference RabbitMqSupervisorBundle

BC break when updating from v1.* to v2.*

If you used custom commands before version 2.0, you need to update them. In most case you can just remove everything after the command name., _(*13)

BC break when updating from v2.* to v3.*

Commands will by default no longer wait for supervisord to complete. If you need this (e. g. to get feedback on errors) use the --wait-for-supervisord option., _(*14)

Usage

Build or rebuild the supervisor and worker configuration and start the daemon:, _(*15)

$ console rabbitmq-supervisor:rebuild

Control the supervisord daemon:, _(*16)

$ console rabbitmq-supervisor:control stop
$ console rabbitmq-supervisor:control start
$ console rabbitmq-supervisor:control restart
$ console rabbitmq-supervisor:control hup