clug

NodeJs Cluster management and logging tool

What does this tool do?

• Simplifies cluster/worker setup
• Automatically revives dead workers
• Supports worker connection affinity (sticky sessions)
• Collects all worker logs in one place
• Automatically formats logs and writes them to a file
• Falls back to single process for debugging
• Monitors memory per worker and restarts worker if limit reached
• Gracefully closes server connections on shutdown or crash

Usage

Basic

var Clug = require('clug');

Clug('./my-script', {
  logPath: '/var/logs/my-script'
});

This will automatically create workers that run my-script.js. By Default it will create as many workers as there are cores on the sytem with a minimum of two (for resiliency).

Sticky sessions

master.js

var Clug = require('clug');

Clug('./my-script', {
  logPath: '/var/logs/my-script',
  sticky: 8001
});

This will automatically route requests to port 8001 to the same worker for every request from the same remote address. Clug automatically handles the intricacies of address sharing for workers so you can code your worker as if it was a master.

For Example:

worker.js

var http = require('http');

server.listen(8001, function(){
  console.info('Server listening on', 8001);
});

This worker listens on the same address as the master so it will be handled by clug.

API

Clug(path, [opts])

• path:String - path to script, used just like require
• opts:Object - options for clug
  • logLevel:String - winston logger log level, defaults to 'debug'
  • logPath:String - directory to put logs into, defaults to root directory of node app
  • workers:Number - number of workers to create, defaults to number of cores with a minimum of two
  • sticky:String|Number|Object|Array - address or addresses to setup sticky sessions for
    • can be port or socket path
    • can be object {host, port, path} (node v0.12+)
    • can be an array of multiple addresses
  • memoryLimit:Number - number of bytes in memory allowed per worker