pm5 (Process Manager) - README

Overview

pm5 is a Python package inspired by pm2, designed to easily manage multiple services and their configurations. It is particularly useful for advanced scenarios where precise control over CPU usage, service monitoring, and automatic restarts are required. This package was created to overcome limitations in managing multiple Celery queues and workers, allowing for more granular control of allocated CPU.

pm5 is a simple tool for developers and system administrators who need advanced control over multiple services and their configurations. Its ability to manage resources, automate restarts, and handle graceful shutdowns makes it an essential tool for complex service management scenarios.

Use Cases

• Managing Multiple Celery Queues: Gain fine-grained control over the number of workers per queue, overcoming limitations of Celery's worker management.
• Service Stack Management: Manage and configure multiple services as a single stack, simplifying the deployment and monitoring process.
• Advanced Resource Management: Allocate CPU resources more effectively across multiple services, ensuring optimal performance and resource utilization.

Features

• Service Management: Start, stop, and monitor multiple services with ease.
• Advanced CPU Management: Control the number of instances per service based on CPU availability.
• Automatic Restarts: Automatically restart services on failure with configurable restart limits.
• Environment Management: Pass specific environment variables to each service.
• Graceful Shutdown: Handle graceful shutdowns and cleanups of services.
• Configuration Flexibility: Easily configure services through a JSON ecosystem file.

Installation

pip install pm5

Configuration

Create an ecosystem configuration file (e.g., ecosystem.config.json) to define the services you want to manage.

Example ecosystem.config.json

{
  "services": [
    {
      "name": "Test Application 1",
      "interpreter": "python3.9",
      "interpreter_args": [],
      "script": "test.py",
      "args": [],
      "instances": 1,
      "wait_ready": true,
      "autorestart": true,
      "max_restarts": 3,
      "env": {
        "ENV_VAR": "value"
      },
      "disabled": false
    }
  ]
}

Configuration Fields

Field	Type	Example	Description
disabled	boolean	false	Enable or disable the service
name	string	Test Application 1	The name of the service used for debugging
interpreter	string	python3.9	The path to the interpreter
interpreter_args	string[]	[]	The args passed to the interpreter
script	string	test.py	The script to call
args	string[]	[]	The args passed to the script
instances	number	1	The number of instances of the script to run
wait_ready	boolean	true	Wait for the service to load before continuing to the next service
autorestart	boolean	true	Automatically restart the service
max_restarts	number	3	The number of times to autorestart the service if failure before exiting
env	object	{}	An object of environment key values that should be passed to the script

Usage

Help

❯ pm5 --help
usage: pm5 [-h] {start,stop} ...

Like pm2 but without node.js.

positional arguments:
  {start,stop}
    start       Start the process manager daemon
    stop        Stop the process manager daemon

optional arguments:
  -h, --help    show this help message and exit

Start the stack as a daemon

pm5 start

Stop the daemonized stack

pm5 stop

Start the app without daemonizing

pm5 start --debug