Reliable, distributed cron.
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
bin
lib
test
web
.codeclimate.yml
.eslintignore
.eslintrc
.gitignore
.rubocop.yml
.travis.yml
CHANGELOG.md
Gemfile
LICENSE.txt
README.md
Rakefile
zhong.gemspec

README.md

Zhong Build Status Code Climate Test Coverage Gem Version

Useful, reliable distributed cron. Tired of your cron-like scheduler running key jobs twice? Would you like to be able to run your cron server on multiple machines and have it "just work"? Have we got the gem for you.

Zhong uses Redis to acquire exclusive locks on jobs, as well as recording when they last ran. This means that you can rest easy at night, knowing that your customers are getting their monthly Goat Fancy magazine subscriptions and you are rolling around in your piles of money without a care in the world.

🍊 Battle-tested at Instacart

Installation

Add this line to your application’s Gemfile:

gem 'zhong'

Usage

Zhong schedule

Create a definition file, let's call it zhong.rb:

Zhong.redis = Redis.new(url: ENV["ZHONG_REDIS_URL"])

Zhong.schedule do
  category "stuff" do
    every 5.seconds, "foo" do
      puts "foo"
    end

    every(1.minute, "running biz at 26th and 27th minute", at: ["**:26", "**:27"]) { puts "biz" }
    every(1.week, "running baz on mon and wed", at: ["mon 22:45", "wed 23:13"]) { puts "baz" }
    every(10.seconds, "boom every 10 seconds") { raise "fail" }
  end

  category "clutter" do
    every(1.second, "compute", if: -> (t) { t.wday == 3 && rand < 0.5 }) do
      puts "something happened on wednesday, maybe"
    end
  end

  # note: callbacks that explicitly false will cause event to not run
  on(:before_tick) do
    puts "ding"
    true
  end

  on(:after_tick) do
    puts "dong"
  end

  on(:before_run) do |job, time|
    puts "running #{job}"
    true # can conditionally run a specific job
  end

  on(:after_run) do |job, time, ran|
    puts "#{job} ran?: #{ran}"
  end

  on(:before_disable) do |job|
    puts "#{job} is going to be disabled"
  end

  on(:after_disable) do |job|
    puts "#{job} disabled"
  end

  on(:before_enable) do |job|
    puts "#{job} is going to be enabled"
  end

  on(:after_enable) do |job|
    puts "#{job} enabled"
  end

  error_handler do |e, job|
    puts "dang, #{job} messed up: #{e}"
  end
end

This file only describes what should be the schedule. Nothing will be executed until we actually run

Zhong.start

after describing the Zhong schedule.

Zhong cron process

You can run the cron process that will execute your code from the definitions in the zhong.rb file by running:

zhong zhong.rb

Web UI

Zhong comes with a web application that can display jobs, their last run and enable/disable them.

This is a Sinatra application that requires at least v2.0.0. You can add to your Gemfile

gem 'sinatra', "~>2.0"

It can be protected by HTTP basic authentication by setting the following environment variables:

  • ZHONG_WEB_USERNAME: the username
  • ZHONG_WEB_PASSWORD: the password

You'll need to load the Zhong schedule to be able to see jobs in the web UI, typically by requiring your zhong.rb definition file.

Rails

Load the Zhong schedule by creating an initializer at config/initializers/zhong.rb, with the following content:

require "#{Rails.root}/zhong.rb"

Add the following to your config/routes.rb:

require 'zhong/web'

Rails.application.routes.draw do
  # Other routes here...

  mount Zhong::Web, at: "/zhong"
end

History

View the changelog.

Contributing

Everyone is encouraged to help improve this project. Here are a few ways you can help: