Backburner

Backburner is a beanstalkd-powered job queue that can handle a very high volume of jobs. You create background jobs and place them on multiple work queues to be processed later.

Processing background jobs reliably has never been easier than with beanstalkd and Backburner. This gem works with any ruby-based web framework, but is especially suited for use with Sinatra, Padrino and Rails.

If you want to use beanstalk for your job processing, consider using Backburner. Backburner is heavily inspired by Resque and DelayedJob. Backburner stores all jobs as simple JSON message payloads. Backburner can be a persistent queue when the beanstalk persistence mode is enabled. It supports multiple queues, priorities, delays, and timeouts.

Why Backburner?

Backburner is well tested and has a familiar, no-nonsense approach to job processing, but that is of secondary importance. Let's face it, there are a lot of options for background job processing. DelayedJob, and Resque are the first that come to mind immediately. So, how do we make sense of which one to use? And why use Backburner over other alternatives?

The key to understanding the differences lies in understanding the different projects and protocols that power these popular queue libraries under the hood. Every job queue requires a queue store that jobs are put into and pulled out of. In the case of Resque, jobs are processed through Redis, a persistent key-value store. In the case of DelayedJob, jobs are processed through ActiveRecord and a database such as PostgreSQL.

The work queue underlying these gems tells you infinitely more about the differences than anything else. Beanstalk is probably the best solution for job queues available today for many reasons. The real question then is... "Why Beanstalk?".

Why Beanstalk?

Illya has an excellent blog post Scalable Work Queues with Beanstalk and Adam Wiggins posted an excellent comparison.

You will quickly see that beanstalkd is an underrated but incredible project that is extremely well-suited as a job queue. Significantly better suited for this task than Redis or a database. Beanstalk is a simple, and a very fast work queue service rolled into a single binary - it is the memcached of work queues. Originally built to power the backend for the 'Causes' Facebook app, it is a mature and production ready open source project. PostRank uses beanstalk to reliably process millions of jobs a day.

A single instance of Beanstalk is perfectly capable of handling thousands of jobs a second (or more, depending on your job size) because it is an in-memory, event-driven system. Powered by libevent under the hood, it requires zero setup (launch and forget, à la memcached), optional log based persistence, an easily parsed ASCII protocol, and a rich set of tools for job management that go well beyond a simple FIFO work queue.

Beanstalk supports the following features natively, out of the box, without any questions asked:

• Parallel Queues - Supports multiple work queues created on demand.
• Reliable - Beanstalk’s reserve, work, delete cycle ensures reliable processing.
• Scheduling - Delay enqueuing jobs by a specified interval to schedule processing later.
• Fast - Processes thousands of jobs per second; significantly faster than alternatives.
• Priorities - Specify priority so important jobs can be processed quickly.
• Persistence - Jobs are stored in memory for speed, but logged to disk for safe keeping.
• Federation - Horizontal scalability provided through federation by the client.
• Error Handling - Bury any job which causes an error for later debugging and inspection.

Keep in mind that these features are supported out of the box with beanstalk and require no special code within this gem to support. In the end, beanstalk is the ideal job queue while also being ridiculously easy to install and setup.

Installation

First, you probably want to install beanstalkd, which powers the job queues. Depending on your platform, this should be as simple as (for Ubuntu):

$ sudo apt-get install beanstalkd

Add this line to your application's Gemfile:

gem 'backburner'

And then execute:

$ bundle

Or install it yourself as:

$ gem install backburner

Configuration

Backburner is extremely simple to setup. Just configure basic settings for backburner:

Backburner.configure do |config|
  config.beanstalk_url = "beanstalk://127.0.0.1"
  config.tube_namespace = "some.app.production"
  config.on_error = lambda { |e| puts e }
  config.default_priority = 65536
  config.respond_timeout = 120
end

Usage

Backburner allows you to create jobs and place them on a beanstalk queue, and later pull those jobs off the queue and process them asynchronously.

Enqueuing Jobs

At the core, Backburner is about jobs that can be processed. Jobs are simple ruby objects with a method defined named perform.

Any object which responds to perform can be queued as a job. Job objects are queued as JSON to be later processed by a task runner. Here's an example:

class NewsletterJob
  include Backburner::Queue
  queue "newsletter"  # defaults to 'newsletter-job'
  queue_priority 1000 # most urgent priority is 0

  def self.perform(email, body)
    NewsletterMailer.deliver_text_to_email(email, body)
  end
end

Notice that you can include the optional Backburner::Queue module so you can specify a queue name for this job. Jobs can be enqueued with:

Backburner.enqueue NewsletterJob, '[email protected]', 'lorem ipsum...'

Backburner.enqueue accepts first a ruby object that supports perform and then a series of parameters to that object's perform method. The queue name used by default is the normalized class name (i.e {namespace}.newsletter-job) if not otherwise specified.

Simple Async Jobs

In addition to defining custom jobs, a job can also be enqueued by invoking the async method on any object which includes Backburner::Performable. Async enqueuing works for both instance and class methods on any performable object.

class User
  include Backburner::Performable
  queue "user-jobs"  # defaults to 'user'
  queue_priority 500 # most urgent priority is 0

  def activate(device_id)
    @device = Device.find(device_id)
    # ...
  end

  def self.reset_password(user_id)
    # ...
  end
end

# Async works for instance methods on a persisted model
@user = User.first
@user.async(:ttr => 100, :queue => "activate").activate(@device.id)
# ..as well as for class methods
User.async(:pri => 100, :delay => 10.seconds).reset_password(@user.id)

This will automatically enqueue a job for that user record that will run activate with the specified argument. Note that you can set the queue name and queue priority at the class level and you are also able to pass pri, ttr, delay and queue directly as options into async. The queue name used by default is the normalized class name (i.e {namespace}.user) if not otherwise specified.

Working Jobs

Backburner workers are processes that run forever handling jobs that get reserved. Starting a worker in ruby code is simple:

Backburner.work

This will process jobs in all queues but you can also restrict processing to specific queues:

Backburner.work('newsletter_sender')

The Backburner worker also exists as a rake task:

require 'backburner/tasks'

so you can run:

$ QUEUES=newsletter-sender,push-message rake backburner:work

You can also run the backburner binary for a convenient worker:

bundle exec backburner newsletter-sender,push-message -d -P /var/run/backburner.pid -l /var/log/backburner.log

This will daemonize the worker and store the pid and logs automatically.

Default Queues

Workers can be easily restricted to processing only a specific set of queues as shown above. However, if you want a worker to process all queues instead, then you can leave the queue list blank.

When you execute a worker without queues specified, any queue for a known job queue class with include Backburner::Queue will be processed. To access the list of known queue classes, you can use:

Backburner::Worker.known_queue_classes
# => [NewsletterJob, SomeOtherJob]

Dynamic queues created by passing queue options will not be processed by a default worker. For this reason, you may want to take control over the default list of queues processed when none are specified. To do this, you can use the default_queues class method:

Backburner.default_queues.concat(["foo", "bar"])

This will ensure that the foo and bar queues are processed by default. You can also add job queue names:

Backburner.default_queues << NewsletterJob.queue

The default_queues stores the specific list of queues that should be processed by default by a worker.

Failures

You can setup the error handler for jobs using configure:

Backburner.configure do |config|
  config.on_error = lambda { |ex| Airbrake.notify(ex) }
end

Now all beanstalk queue errors will show up on airbrake. If a job fails in beanstalk, the job is automatically buried and must be 'kicked' later.

Logging

Right now, all logging happens to standard out and can be piped to a file or any other output manually. More on logging coming later.

Web Front-end

Be sure to check out the Sinatra-powered project beanstalkd_view by denniskuczynski which provides an excellent overview of the tubes and jobs processed by your beanstalk workers. An excellent addition to your Backburner setup.

Workers in Production

Once you have Backburner setup in your application, starting workers is really easy. Once beanstalkd is installed, your best bet is to use the built-in rake task that comes with Backburner. Simply add the task to your Rakefile:

# Rakefile
require 'backburner/tasks'

and then you can start the rake task with:

$ rake backburner:work
$ QUEUES=newsletter-sender,push-message rake backburner:work

The best way to deploy these rake tasks is using a monitoring library. We suggest God which watches processes and ensures their stability. A simple God recipe for Backburner can be found in examples/god.

Acknowledgements

• Nathan Esquenazi - Project maintainer
• Kristen Tucker - Coming up with the gem name
• Tim Lee, Josh Hull, Nico Taing - Helping me work through the idea
• Miso - Open-source friendly place to work

Contributing

1. Fork it
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Added some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create new Pull Request

References

The code in this project has been made in light of a few excellent projects:

• DelayedJob
• Resque
• Stalker

Thanks to these projects for inspiration and certain design and implementation decisions.

Links

• Code: git clone git://github.com/nesquena/backburner.git
• Home: http://github.com/nesquena/backburner
• Docs: http://rdoc.info/github/nesquena/backburner/master/frames
• Bugs: http://github.com/nesquena/backburner/issues
• Gems: http://gemcutter.org/gems/backburner