Project

boss_queue

0.0
No commit activity in last 3 years
No release in over 3 years
A fault tolerant job queue built around Amazon SQS & DynamoDB
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
 Dependencies

Development

Runtime

>= 1.19.0
 Project Readme

boss_queue

A fault tolerant job queue built around Amazon SQS & DynamoDB

Setup

In your Gemfile:

gem 'boss_queue'

boss_queue uses an Amazon SQS queue and a Amazon DynamoDB table for each environment (production, staging, test). To set these up in Rails do:

$ rails c

AWS.config(:access_key_id => <access_key_id>,
           :secret_access_key => <secret_access_key>)

queue = BossQueue.new

BossQueue.environment = 'development'
queue.create_table
queue.create_queue

BossQueue.environment = 'staging'
queue.create_table
queue.create_queue

# BossQueue.create_table(read_capacity, write_capacity)
# One read capacity unit = two eventually consistent reads per second, for items up 4 KB in size.
# One write capacity unit = one write per second, for items up to 1 KB in size.

BossQueue.environment = 'production'
queue.create_table(50, 10)
queue.create_queue

# you can also have customized queue names for separate job queues
# (only create the queue again...a single DynamoDB table serves all the queues)
queue = BossQueue.new(:queue => 'emails')
queue.create_queue
queue = BossQueue.new(:queue => 'image_processing')
queue.create_queue

Alternatively, in each of the respective environments, do:

$ rails c

AWS.config(:access_key_id => <access_key_id>,
           :secret_access_key => <secret_access_key>)
queue = BossQueue.new

# environment does not need to be set because it is taken from Rails.env
queue.create_table(50, 10)
queue.create_queue

Or these could be put into a migration.

Usage

myobject = MyClass.new
# default failure action is 'retry' which retries up to four times
queue = BossQueue.new(:failure_action => 'none', :queue => 'emails')

# or we can set up a callback method to be called on the enqueued class / object
# when there is a failure
queue = BossQueue.new(:failure_action => 'callback', :failure_callback => :method_to_execute_on_failure, :queue => 'emails')

# by default, BossQueue will record a failed job if the target cannot be found
# when the background job runs; the job can be deleted instead by including a
# :delete_if_target_missing => true option
queue = BossQueue.new(:delete_if_target_missing => true, :queue => 'emails')

# can enqueue instance methods (assumes that objects have an id and a #find(id) method)
queue.enqueue(myobject, :method_to_execute, arg1, arg2)
# enqueue with a delay of up to 900 seconds (15 minutes)
queue.enqueue_with_delay(60, myobject, :method_to_execute, arg1, arg2, arg3)

# can enqueue class methods
queue.enqueue(MyClass, :method_to_execute)
queue.enqueue_with_delay(60, MyClass, :method_to_execute, arg1, arg2)

# work returns true if a job was pulled from the queue, false otherwise
queue.work

# failures are left in the DynamoDB table with the failed boolean set to true

BossQueue does not at present have a daemon component like Sidekiq and Resque do.

Future Work

Create some mechanism for viewing failed jobs (and perhaps queued jobs...they are all in the same table)