For those (like me) who have/had no idea what background jobs are, let’s start with that (though I’m not even sure I know exactly what they are, to be fair). I guess the idea is that, sometimes things fail, and when things fail, you don’t want everything to fail. Like lightbulbs in series vs lightbulbs in parallel. When they’re in series, if one bulb goes out, they all go out. But if they’re in parallel, when one dies, the others don’t go down with it. So, in your app, if you’re sending out all these emails, and one of them fails, you don’t want all subsequent emails to fail! You want the rest to keep truckin, right?

Hence libraries like delayed_job and resque, which let you add jobs to queues and use separate workers to take jobs from the queues and perform them. If one worker fails, it’s fine, cause another worker will come along and finish off the queue!

So now that we have a preliminary idea of what background jobs are, let’s start. Resque is a redis-backed library, so you’ll need to install redis for this to work. You can install it with brew $ brew install redis or you can download the latest tarball. Then you’ll need to run redis in the background with $ redis-server.

Next, add the following two gems to your gemfile and bundle:

gem 'resque'
gem 'resque-scheduler'

Rails uses ActiveJob as its main job infrastructure, so we can continue using ActiveJob and not have to change much code, but tell it to use Resque as its queueing backend.

# config/initializers/active_job.rb
ActiveJob::Base.queue_adapter = :resque

Next, we’ll need to tell our app where our redis server is located, which will by default run on port 6379

# config/initializers/resque.rb
Resque.redis = Redis.new(url: 'redis://localhost:6379')

The next part is optional, but helpful. If you add the following lines to your routes file, you’ll be able to go to ‘/resque’ and see all your queues, jobs, and failing jobs on a nice little interface. You can even retry jobs from there.

require 'resque/server'
Rails.application.routes.draw do
  mount Resque::Server.new, at: '/resque'
  # The rest of your routes...

Alright, last part of the setup: you’ll need to load the rake tasks for resque.

require 'resque/tasks'
require 'resque/scheduler/tasks'
namespace :resque do
  task setup: :environment do
     require 'resque'
     require 'resque-scheduler'
  end
end

If you run $ rake -T resque you’ll be able to see all tasks that were loaded by resque

Now, let’s start adding our jobs to queues. If you’re already sending some emails in your rails app, then all we have to do is tell the Mailer objects not to deliver now, but to “deliver later” which will add those jobs to the “mailer” queue instead of performing them immediately.

MyMailer.send_cool_email(user: current_user).deliver_later
# instead of
MyMailer.send_cool_email(user: current_user).deliver_now

Now, if you go into your rails console, you should be able to check that there are jobs in the queue.

MyMailer.send_cool_email(user: User.first).deliver_later
Resque.size('mailers') # 1

To start some workers to perform these jobs, you can run the rake task with a specified queue, or set QUEUE=* to perform all jobs from any queue.

$ rake resque:work QUEUE=mailers

And that’s pretty much it! One more thing though, if you’re using Mongoid with resque, you’ll need to add include GlobalID::Identification to any model you’re passing to the Mailer methods (in these examples, the User model). If you wanna know why, you can read about it here.