/delayed_job

Database based asynchronous priority queue system -- Extracted from Shopify

Primary LanguageRubyMIT LicenseMIT

If you're viewing this at https://github.com/collectiveidea/delayed_job, you're reading the documentation for the master branch. View documentation for the latest release (4.1.11).

Delayed::Job

Gem Version CI Code Climate Coverage Status

Delayed::Job (or DJ) encapsulates the common pattern of asynchronously executing longer tasks in the background.

It is a direct extraction from Shopify where the job table is responsible for a multitude of core tasks. Amongst those tasks are:

  • sending massive newsletters
  • image resizing
  • http downloads
  • updating smart collections
  • updating solr, our search server, after product changes
  • batch imports
  • spam checks

Follow us on Twitter to get updates and notices about new releases.

Installation

delayed_job 3.0.0 only supports Rails 3.0+.

delayed_job supports multiple backends for storing the job queue. See the wiki for other backends.

If you plan to use delayed_job with Active Record, add delayed_job_active_record to your Gemfile.

gem 'delayed_job_active_record'

If you plan to use delayed_job with Mongoid, add delayed_job_mongoid to your Gemfile.

gem 'delayed_job_mongoid'

Run bundle install to install the backend and delayed_job gems.

The Active Record backend requires a jobs table. You can create that table by running the following command:

rails generate delayed_job:active_record
rake db:migrate

For Rails 4.2+, see below

Development

In development mode, if you are using Rails 3.1+, your application code will automatically reload every 100 jobs or when the queue finishes. You no longer need to restart Delayed Job every time you update your code in development.

Active Job

In Rails 4.2+, set the queue_adapter in config/application.rb

config.active_job.queue_adapter = :delayed_job

See the rails guide for more details.

Rails 4.x

If you are using the protected_attributes gem, it must appear before delayed_job in your gemfile. If your jobs are failing with:

 ActiveRecord::StatementInvalid: PG::NotNullViolation: ERROR:  null value in column "handler" violates not-null constraint

then this is the fix you're looking for.

Upgrading from 2.x to 3.0.0 on Active Record

Delayed Job 3.0.0 introduces a new column to the delayed_jobs table.

If you're upgrading from Delayed Job 2.x, run the upgrade generator to create a migration to add the column.

rails generate delayed_job:upgrade
rake db:migrate

Queuing Jobs

Call .delay.method(params) on any object and it will be processed in the background.

# without delayed_job
@user.activate!(@device)

# with delayed_job
@user.delay.activate!(@device)

If a method should always be run in the background, you can call #handle_asynchronously after the method declaration:

class Device
  def deliver
    # long running method
  end
  handle_asynchronously :deliver
end

device = Device.new
device.deliver

Parameters

#handle_asynchronously and #delay take these parameters:

  • :priority (number): lower numbers run first; default is 0 but can be reconfigured (see below)
  • :run_at (Time): run the job after this time (probably in the future)
  • :queue (string): named queue to put this job in, an alternative to priorities (see below)

These params can be Proc objects, allowing call-time evaluation of the value.

For example:

class LongTasks
  def send_mailer
    # Some other code
  end
  handle_asynchronously :send_mailer, :priority => 20

  def in_the_future
    # Some other code
  end
  # 5.minutes.from_now will be evaluated when in_the_future is called
  handle_asynchronously :in_the_future, :run_at => Proc.new { 5.minutes.from_now }

  def self.when_to_run
    2.hours.from_now
  end

  class << self
    def call_a_class_method
      # Some other code
    end
    handle_asynchronously :call_a_class_method, :run_at => Proc.new { when_to_run }
  end

  attr_reader :how_important

  def call_an_instance_method
    # Some other code
  end
  handle_asynchronously :call_an_instance_method, :priority => Proc.new {|i| i.how_important }
end

If you ever want to call a handle_asynchronously'd method without Delayed Job, for instance while debugging something at the console, just add _without_delay to the method name. For instance, if your original method was foo, then call foo_without_delay.

Rails Mailers

Delayed Job uses special syntax for Rails Mailers. Do not call the .deliver method when using .delay.

# without delayed_job
Notifier.signup(@user).deliver

# with delayed_job
Notifier.delay.signup(@user)

# delayed_job running at a specific time
Notifier.delay(run_at: 5.minutes.from_now).signup(@user)

# when using parameters, the .with method must be called before the .delay method
Notifier.with(foo: 1, bar: 2).delay.signup(@user)

You may also wish to consider using Active Job with Action Mailer which provides convenient .deliver_later syntax that forwards to Delayed Job under-the-hood.

Named Queues

DJ 3 introduces Resque-style named queues while still retaining DJ-style priority. The goal is to provide a system for grouping tasks to be worked by separate pools of workers, which may be scaled and controlled individually.

Jobs can be assigned to a queue by setting the queue option:

object.delay(:queue => 'tracking').method

Delayed::Job.enqueue job, :queue => 'tracking'

handle_asynchronously :tweet_later, :queue => 'tweets'

You can configure default priorities for named queues:

Delayed::Worker.queue_attributes = {
  high_priority: { priority: -10 },
  low_priority: { priority: 10 }
}

Configured queue priorities can be overriden by passing priority to the delay method

object.delay(:queue => 'high_priority', priority: 0).method

You can start processes to only work certain queues with the queue and queues options defined below. Processes started without specifying a queue will run jobs from any queue. To effectively have a process that runs jobs where a queue is not specified, set a default queue name with Delayed::Worker.default_queue_name and have the processes run that queue.

Running Jobs

script/delayed_job can be used to manage a background process which will start working off jobs.

To do so, add gem "daemons" to your Gemfile and make sure you've run rails generate delayed_job.

You can then do the following:

RAILS_ENV=production script/delayed_job start
RAILS_ENV=production script/delayed_job stop

# Runs two workers in separate processes.
RAILS_ENV=production script/delayed_job -n 2 start
RAILS_ENV=production script/delayed_job stop

# Set the --queue or --queues option to work from a particular queue.
RAILS_ENV=production script/delayed_job --queue=tracking start
RAILS_ENV=production script/delayed_job --queues=mailers,tasks start

# Use the --pool option to specify a worker pool. You can use this option multiple times to start different numbers of workers for different queues.
# The following command will start 1 worker for the tracking queue,
# 2 workers for the mailers and tasks queues, and 2 workers for any jobs:
RAILS_ENV=production script/delayed_job --pool=tracking --pool=mailers,tasks:2 --pool=*:2 start

# Runs all available jobs and then exits
RAILS_ENV=production script/delayed_job start --exit-on-complete
# or to run in the foreground
RAILS_ENV=production script/delayed_job run --exit-on-complete

Rails 4: replace script/delayed_job with bin/delayed_job

Workers can be running on any computer, as long as they have access to the database and their clock is in sync. Keep in mind that each worker will check the database at least every 5 seconds.

You can also invoke rake jobs:work which will start working off jobs. You can cancel the rake task with CTRL-C.

If you want to just run all available jobs and exit you can use rake jobs:workoff

Work off queues by setting the QUEUE or QUEUES environment variable.

QUEUE=tracking rake jobs:work
QUEUES=mailers,tasks rake jobs:work

Restarting delayed_job

The following syntax will restart delayed jobs:

RAILS_ENV=production script/delayed_job restart

To restart multiple delayed_job workers:

RAILS_ENV=production script/delayed_job -n2 restart

Rails 4: replace script/delayed_job with bin/delayed_job

Custom Jobs

Jobs are simple ruby objects with a method called perform. Any object which responds to perform can be stuffed into the jobs table. Job objects are serialized to yaml so that they can later be resurrected by the job runner.

NewsletterJob = Struct.new(:text, :emails) do
  def perform
    emails.each { |e| NewsletterMailer.deliver_text_to_email(text, e) }
  end
end

Delayed::Job.enqueue NewsletterJob.new('lorem ipsum...', Customers.pluck(:email))

To set a per-job max attempts that overrides the Delayed::Worker.max_attempts you can define a max_attempts method on the job

NewsletterJob = Struct.new(:text, :emails) do
  def perform
    emails.each { |e| NewsletterMailer.deliver_text_to_email(text, e) }
  end

  def max_attempts
    3
  end
end

To set a per-job max run time that overrides the Delayed::Worker.max_run_time you can define a max_run_time method on the job

NOTE: this can ONLY be used to set a max_run_time that is lower than Delayed::Worker.max_run_time. Otherwise the lock on the job would expire and another worker would start the working on the in progress job.

NewsletterJob = Struct.new(:text, :emails) do
  def perform
    emails.each { |e| NewsletterMailer.deliver_text_to_email(text, e) }
  end

  def max_run_time
    120 # seconds
  end
end

To set a per-job default for destroying failed jobs that overrides the Delayed::Worker.destroy_failed_jobs you can define a destroy_failed_jobs? method on the job

NewsletterJob = Struct.new(:text, :emails) do
  def perform
    emails.each { |e| NewsletterMailer.deliver_text_to_email(text, e) }
  end

  def destroy_failed_jobs?
    false
  end
end

To set a default queue name for a custom job that overrides Delayed::Worker.default_queue_name, you can define a queue_name method on the job

NewsletterJob = Struct.new(:text, :emails) do
  def perform
    emails.each { |e| NewsletterMailer.deliver_text_to_email(text, e) }
  end

  def queue_name
    'newsletter_queue'
  end
end

On error, the job is scheduled again in 5 seconds + N ** 4, where N is the number of attempts. You can define your own reschedule_at method to override this default behavior.

NewsletterJob = Struct.new(:text, :emails) do
  def perform
    emails.each { |e| NewsletterMailer.deliver_text_to_email(text, e) }
  end

  def reschedule_at(current_time, attempts)
    current_time + 5.seconds
  end
end

Hooks

You can define hooks on your job that will be called at different stages in the process:

NOTE: If you are using ActiveJob these hooks are not available to your jobs. You will need to use ActiveJob's callbacks. You can find details here https://guides.rubyonrails.org/active_job_basics.html#callbacks

class ParanoidNewsletterJob < NewsletterJob
  def enqueue(job)
    record_stat 'newsletter_job/enqueue'
  end

  def perform
    emails.each { |e| NewsletterMailer.deliver_text_to_email(text, e) }
  end

  def before(job)
    record_stat 'newsletter_job/start'
  end

  def after(job)
    record_stat 'newsletter_job/after'
  end

  def success(job)
    record_stat 'newsletter_job/success'
  end

  def error(job, exception)
    Airbrake.notify(exception)
  end

  def failure(job)
    page_sysadmin_in_the_middle_of_the_night
  end
end

Gory Details

The library revolves around a delayed_jobs table which looks as follows:

create_table :delayed_jobs, :force => true do |table|
  table.integer  :priority, :default => 0      # Allows some jobs to jump to the front of the queue
  table.integer  :attempts, :default => 0      # Provides for retries, but still fail eventually.
  table.text     :handler                      # YAML-encoded string of the object that will do work
  table.text     :last_error                   # reason for last failure (See Note below)
  table.datetime :run_at                       # When to run. Could be Time.zone.now for immediately, or sometime in the future.
  table.datetime :locked_at                    # Set when a client is working on this object
  table.datetime :failed_at                    # Set when all retries have failed (actually, by default, the record is deleted instead)
  table.string   :locked_by                    # Who is working on this object (if locked)
  table.string   :queue                        # The name of the queue this job is in
  table.timestamps
end

On error, the job is scheduled again in 5 seconds + N ** 4, where N is the number of attempts or using the job's defined reschedule_at method.

The default Worker.max_attempts is 25. After this, the job is either deleted (default), or left in the database with "failed_at" set. With the default of 25 attempts, the last retry will be 20 days later, with the last interval being almost 100 hours.

The default Worker.max_run_time is 4.hours. If your job takes longer than that, another computer could pick it up. It's up to you to make sure your job doesn't exceed this time. You should set this to the longest time you think the job could take.

By default, it will delete failed jobs (and it always deletes successful jobs). If you want to keep failed jobs, set Delayed::Worker.destroy_failed_jobs = false. The failed jobs will be marked with non-null failed_at.

By default all jobs are scheduled with priority = 0, which is top priority. You can change this by setting Delayed::Worker.default_priority to something else. Lower numbers have higher priority.

The default behavior is to read 5 jobs from the queue when finding an available job. You can configure this by setting Delayed::Worker.read_ahead.

By default all jobs will be queued without a named queue. A default named queue can be specified by using Delayed::Worker.default_queue_name.

If no jobs are found, the worker sleeps for the amount of time specified by the sleep delay option. Set Delayed::Worker.sleep_delay = 60 for a 60 second sleep time.

It is possible to disable delayed jobs for testing purposes. Set Delayed::Worker.delay_jobs = false to execute all jobs realtime.

Or Delayed::Worker.delay_jobs can be a Proc that decides whether to execute jobs inline on a per-job basis:

Delayed::Worker.delay_jobs = ->(job) {
  job.queue != 'inline'
}

You may need to raise exceptions on SIGTERM signals, Delayed::Worker.raise_signal_exceptions = :term will cause the worker to raise a SignalException causing the running job to abort and be unlocked, which makes the job available to other workers. The default for this option is false.

Here is an example of changing job parameters in Rails:

# config/initializers/delayed_job_config.rb
Delayed::Worker.destroy_failed_jobs = false
Delayed::Worker.sleep_delay = 60
Delayed::Worker.max_attempts = 3
Delayed::Worker.max_run_time = 5.minutes
Delayed::Worker.read_ahead = 10
Delayed::Worker.default_queue_name = 'default'
Delayed::Worker.delay_jobs = !Rails.env.test?
Delayed::Worker.raise_signal_exceptions = :term
Delayed::Worker.logger = Logger.new(File.join(Rails.root, 'log', 'delayed_job.log'))

Cleaning up

You can invoke rake jobs:clear to delete all jobs in the queue.

Having problems?

Good places to get help are: