huangdijia / laravel-horizon-restart
Horizon Restart for Laravel.
Fund package maintenance!
huangdijia
hdj.me/sponsors
Installs: 100 677
Dependents: 0
Suggesters: 0
Security: 0
Stars: 23
Watchers: 1
Forks: 2
Open Issues: 0
pkg:composer/huangdijia/laravel-horizon-restart
Requires
- php: >=8.1
- illuminate/bus: ^10.0|^11.0|^12.0
- illuminate/console: ^10.0|^11.0|^12.0
- illuminate/support: ^10.0|^11.0|^12.0
- laravel/horizon: ^5.0|^6.0
Requires (Dev)
- huangdijia/php-coding-standard: ^1.2
- laravel/framework: ^10.0|^11.0|^12.0
- pestphp/pest: ^2.0|^3.7
- pestphp/pest-plugin-laravel: ^2.0|^3.1
- phpstan/phpstan: ^1.0|^2.1
README
A Laravel package that extends Laravel Horizon with the ability to restart supervisors across multiple servers, similar to php artisan queue:restart but specifically designed for Horizon workers.
Features
- π Restart all Horizon supervisors across multiple servers with a single command
- π Graceful termination of workers similar to
queue:restart - π¦ Automatic service provider registration
- π― Works seamlessly with Horizon's queue distribution
- β‘ Zero downtime deployments for Horizon workers
Requirements
- PHP 8.2 or higher
- Laravel 11.x or 12.x
- Laravel Horizon 5.x or 6.x
Installation
Install the package via Composer:
composer require huangdijia/laravel-horizon-restart
The service provider will be automatically registered.
Usage
Basic Usage
To restart all Horizon supervisors across all servers:
php artisan horizon:restart
This command will:
- Discover all running Horizon supervisors
- Dispatch a restart job to each supervisor's queue
- Gracefully terminate each supervisor (similar to
horizon:terminate) - Horizon will automatically restart the supervisors
How It Works
When you run horizon:restart, the package:
- Discovers Supervisors: Queries the Horizon master supervisor repository to find all active supervisors
- Dispatches Jobs: Creates a
HorizonRestartJobfor each supervisor and dispatches it to that supervisor's queue - Executes Termination: Each job runs on its respective server and calls
horizon:terminatelocally - Auto Restart: Horizon's process monitoring automatically restarts terminated supervisors
This approach ensures that:
- All servers receive the restart signal
- Workers finish their current jobs before terminating
- No jobs are lost during the restart process
- Each server restarts independently
Use Cases
This package is particularly useful for:
- Multi-Server Deployments: When you have Horizon running on multiple servers and need to restart all of them
- Code Deployments: After deploying new code, restart all workers across all servers
- Configuration Changes: When Horizon configuration changes and all workers need to reload
- Memory Management: Periodic restarts to clear memory leaks or reset worker state
Configuration
The package automatically configures a special supervisor for handling restart jobs. No manual configuration is required.
The restart supervisor is automatically configured with:
- Connection: Uses the same connection as your first Horizon environment
- Queue: Uses Horizon's internal queue name
- Processes: 1 worker to handle restart jobs
- Tries: 3 attempts per job
Comparison with Other Methods
| Method | Single Server | Multiple Servers | Graceful |
|---|---|---|---|
horizon:terminate |
β | β | β |
queue:restart |
β | β | β |
horizon:restart (this package) |
β | β | β |
Development
Testing
Run the test suite:
composer test
Code Style
Fix code style issues:
composer cs-fix
Static Analysis
Run static analysis:
composer analyse
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
Security
If you discover any security related issues, please email huangdijia@gmail.com instead of using the issue tracker.
License
The MIT License (MIT). Please see License File for more information.