Dispatch::Rider
Dispatch rider is a pub/sub kind of library that allows you to publish a message to a notification system (like Amazon SNS) and then you can subscribe to the channels that you subscribed to and start handling the messages. Essentially making it a multi-service background job system.
Build status
Installation
Add this line to your application's Gemfile:
gem 'dispatch-rider'
And then execute:
$ bundle
Or install it yourself as:
$ gem install dispatch-rider
If you are using DispatchRider with rails, run the installer:
$ rails generate dispatch_rider:install
Usage
Publisher
Setting up a publisher is simple.
Hash Based Configuration
All configuration can be loaded from a hash instead of being done like the examples below. (currently only implemented for the publisher)
Global configuration
You can set the global configuration using either a hash:
DispatchRider::Publisher.configure({
notification_services: {
file_system: {}
},
destinations: {
file_foo: {
service: :file_system,
channel: :foo,
options: {
path: "test/channel",
}
}
}
})
or a block:
DispatchRider::Publisher.configure do |config|
config.parse({
notification_services: {
file_system: {}
},
destinations: {
file_foo: {
service: :file_system,
channel: :foo,
options: {
path: "test/channel",
}
}
}
})
end
Then anytime you call configure on a new publisher, it will default to global configuration.
DispatchRider::Publisher.new
# is the same as
DispatchRider::Publisher.new(DispatchRider::Publisher.configuration)
Local configuration
Alternatively, you can create your own configuration and load that configuration into your new publisher.
config = DispatchRider::Publisher::Configuration.new({
notification_services: {
file_system: {}
},
destinations: {
file_foo: {
service: :file_system,
channel: :foo,
options: {
path: "test/channel",
}
}
}
})
DispatchRider::Publisher.new(config)
You can load this configuration hash from a YAML file or something, whatever works well for your environment.
The old way ...
To publish using the filesystem register the path where to publish the message files.
publisher = DispatchRider::Publisher.new
publisher.register_notification_service(:file_system)
publisher.register_destination(:local_message_queue, :file_system, :dev_channel, :path => "tmp/news-updates")
publisher.publish(:destinations => :local_message_queue, :message => {
:subject => "read_news",
:body => {"headlines" => [
"April 29, 2013: Rails 4.0.0.rc1 is released.",
"May 14, 2013: Ruby 2.0.0-p195 is released"
]
}})
To publish using AWS::SNS
make sure AWS.config
has been setup.
It's then as easy as providing the configuration details of the topic to the publisher.
publisher = DispatchRider::Publisher.new
publisher.register_notification_service(:aws_sns)
publisher.register_destination(:sns_message_queue, :aws_sns, :dev_channel, {
:account => 777,
:region => 'us-east-1',
:topic => 'RoR'
})
publisher.publish(:destinations => :sns_message_queue, :message => {
:subject => "read_news",
:body => {"headlines" => [
"April 29, 2013: Rails 4.0.0.rc1 is released.",
"May 14, 2013: Ruby 2.0.0-p195 is released"
]
}})
To publish to multiple destinations:
publisher.publish(:destinations => [:local_message_queue, :sns_message_queue], :message => {
:subject => "read_news",
:body => {"headlines" => [
"April 29, 2013: Rails 4.0.0.rc1 is released.",
"May 14, 2013: Ruby 2.0.0-p195 is released"
]
}})
Sample Rails publisher:
# app/publishers/news_update
class NewsPublisher < DispatchRider::Publisher::Base
destinations :sns_message_queue
subject "read_news"
def self.publish(news)
new.publish({"headlines" => news.headlines})
end
end
Subscriber
Configuration
You can configure the subscription side of DispatchRider by using the built in configuration object.
DispatchRider.config do |config|
config.before(:initialize) do
# code to run before initialize
end
config.after(:process) do
# code to run after process
end
# allows you to wrap a callback around the execution of each job
config.around(:dispatch_message) do |job, message|
some_block_around do
job.call
end
end
config.logger = Rails.logger
# Log Formatter
# Text Formatter (default)
# config.log_formatter = DispatchRider::Logging::TextFormatter.new
# JSON Formatter
# config.log_formatter = DispatchRider::Logging::JsonFormatter.new
config.default_retry_timeout = 300
config.error_handler = DispatchRider::DefaultErrorHandler # an object that responds to .call(message, exception)
config.queue_kind = :sqs
config.queue_info = { name: "queue-production" }
config.handler_path = Rails.root + "app/handlers" # path to handler files to be autoloaded
end
Options:
-
logger
: what logger to use to send messages to (responds to the standard ruby Logger protocol), defaults to a new Logger sending messages to STDERR -
log_formatter
: what log formatter to use with logger, defaults to DispatchRider::Logging::TextFormatter.new
Callbacks
Dispatch rider supports injecting callbacks in a few parts of the lifecycle of the process. Each
callback can have hooks plugged into it at before
, after
and around
the execution.
On initialize :initialize
This callback is called when the runner is being initialized.
Block Arguments:
- None
DispatchRider.config do |config|
config.before(:initialize) do
# your code here
end
config.around(:initialize) do |job|
# your code here
job.call
# your code here
end
config.after(:initialize) do
# your code here
end
end
On publish :publish
This callback is called when the message is being published.
Block Arguments:
- message:
[DispatchRider::Message]
-- message that is about to be sent - destinations:
[Array<Symbol>]
-- list of destinations to sent to
DispatchRider.config do |config|
config.before(:publish) do |message:, destinations:|
# your code here
end
config.around(:publish) do |job, message:, destinations:|
# your code here
job.call
# your code here
end
config.after(:publish) do |message:, destinations:|
# your code here
end
end
On process :process
This callback is called when the runner is running its event loop.
Block Arguments:
- None
DispatchRider.config do |config|
config.before(:process) do
# your code here
end
config.around(:process) do |job|
# your code here
job.call
# your code here
end
config.after(:process) do
# your code here
end
end
On dispatch message :dispatch_message
This callback is called when executing a single message.
Block Arguments:
- message
[DispatchRider::Message]
-- the message received fromDispatchRider
queue
DispatchRider.config do |config|
config.before(:dispatch_message) do |message|
# your code here
end
config.around(:dispatch_message) do |job, message|
# your code here
job.call
# your code here
end
config.after(:dispatch_message) do |message|
# your code here
end
end
Manual Setup
To setup a subscriber you'll need message handlers. The handlers are named the same as the message subjects. Each handler may also specify a retry_timeout as shown below. When a job throws an exception it will be put back on the queue in that time period if the queue supports timeouts. If the underlying queue (such as filesystem) does not support retry then this setting is ineffective.
Sample message handler:
# app/handlers/bar_handler
class ReadNews < DispatchRider::Handlers::Base
def process(message_body)
message_body["headlines"].each do |headline|
puts headline
end
end
def retry_timeout
10.minutes
end
end
Timeout & retry handling
If you have a long running job, or if you wish to retry a job later, you may use two methods in your handler class. return_to_queue and extend_timeout.
return_to_queue will retry your item immediately. extend_timeout will tell the queue you wish to hold this item longer.
# app/handlers/foo_handler
class LongRunning < DispatchRider::Handlers::Base
def process(body)
my_loop.each do |item|
#... do some work ...
extend_timeout(1.hour)
end
rescue OutOfResourcesImOutError
return_to_queue #oops! Better give this to somebody else!
end
end
Sample subscriber setup:
subscriber = DispatchRider::Subscriber.new
subscriber.register_queue(:aws_sqs, :name => "news-updates")
subscriber.register_handler(:read_news)
subscriber.setup_demultiplexer(:aws_sqs)
subscriber.process
Sample subscriber dispatch error handling (optional):
# using objects
module ErrorHandler
def self.call(message, exception)
# put your error handling code here
return false # or return true to permanently remove the message
end
end
subscriber.setup_demultiplexer(kind, ErrorHandler)
# using lambdas
error_handler = ->(message, exception) do
# put your error handling code here
return false # or return true to permanently remove the message
end
subscriber.setup_demultiplexer(kind, error_handler)
Airbrake Support
Airbrake is supported out of the box. All you need to do is:
- Install and configure the airbrake gem.
- Use the
DispatchRider::AirbrakeErrorHandler
.
subscriber.setup_demultiplexer(kind, DispatchRider::AirbrakeErrorHandler)
or set it up in the config ...
DispatchRider.config do |config|
config.error_handler = DispatchRider::AirbrakeErrorHandler
end
Scheduling Job
Scheduled jobs currently require ActiveRecord
. Support for other ORM will be added in the future depending on demand.
- Add and run the DB migration below:
class CreateScheduledJobsTable < ActiveRecord::Migration
include DispatchRider::ScheduledJob::Migration
def change
create_scheduled_jobs_table
end
end
- Schedule jobs:
# Use `#publish_later`
class NewsPublisher < DispatchRider::Publisher::Base
destinations :sns_message_queue
subject "read_news"
def self.midnight_publish(news)
new.publish_later("headlines" => news.headlines, at: Date.tomorrow.midnight)
end
end
# Or create a scheduled job manually
DispatchRider::ScheduledJob.create! scheduled_at: Date.tomorrow.midnight,
destinations: [:sns_message_queue],
message: {
subject: "read_news",
body: { "headlines" => news.headlines }
}
- Run scheduled publishing.
# Run once
# Ideally run on a cron. Where the cron is responsible for the publishing
# frequency. Any jobs due at the time this is run by the cron will be
# published.
DispatchRider::ScheduledJob.publish_due_jobs
# Loop. Warning: Loops are blocking. Run this on a separate thread if it's not
# the sole purpose of the app.
# publish every minute
DispatchRider::ScheduledJob.publish_due_jobs every: 1.minute
# publish every half an hour
DispatchRider::ScheduledJob.publish_due_jobs every: 30.minutes
Deployment
See DEPLOYING.md
Contributing
Process
- Fork it
- Create your feature branch (
git checkout -b my-new-feature
) - Commit your changes (
git commit -am 'Add some feature'
) - Push to the branch (
git push origin my-new-feature
) - Create new Pull Request
Licence
Copyright (c) 2015 PayrollHero Pte. Ltd.
MIT License
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.