yes, I know it says Starman on the image
🎶 And I think it's gonna be a long long time 'Till touch down brings me round again to find 🎶
Rocketman is a gem that introduces Pub-Sub mechanism within your Ruby code.
The main goal of Rocketman is not to replace proper message buses like Redis PubSub/Kafka, but rather be a stepping stone. You can read more about the rationale behind the project down below.
As with all Pub-Sub mechanism, this greatly decouples your upstream producer and downstream consumer, allowing for scalability, and easier refactor when you decide to move Pub-Sub to a separate service.
Rocketman also works without Rails.
Add this line to your application's Gemfile:
gem 'rocketman'
And then execute:
$ bundle
Or install it yourself as:
$ gem install rocketman
Rocketman exposes two module, Rocketman::Producer
and Rocketman::Consumer
. They do exactly as what their name implies. All you need to do is include Rocketman::Producer
and extend Rocketman::Consumer
into your code.
Producer exposes one instance method to you: :emit
. :emit
takes in the event name and an optional payload and publishes it to the consumers. There's nothing more you need to do. The producer do not have to know who its consumers are.
class Producer
include Rocketman::Producer
def hello_world
emit :hello, payload: {"one" => 1, "two" => 2}
end
end
Note that Producer emit events with threads that run in a thread pool. The default number of worker is 5, and the workers default to checking the job with a 3 seconds interval. You can tweak these to your liking, refer to the Configuration
section below for more informations.
Consumer exposes a class method, :on_event
. :on_event
takes in the event name, and also an additional block, which gets executed whenever a message is received. If an additional payload
is emitted along with the event, you can get access to it in the form of block argument.
class Consumer
extend Rocketman::Consumer
on_event :hello do |payload|
puts "I've received #{payload} here!"
# => I've received {:payload=>{"one"=>1, "two"=>2}} here!
end
end
Simple isn't it?
If you want to also consume events from external services, you can do that too.
Rocketman has the concept of a Relay
. The cool thing is, Relay
s are just Producer
s that understand how to relay messages from external services (like Redis
) into Rocketman events.
Rocketman ships with a Redis
relay, which you can use it like so (assuming you have Redis installed):
require 'rocketman/relay/redis' # This is not required by default, so you need to explicitly require it.
Rocketman::Relay::Redis.new.start(Redis.new)
NOTE: You should always pass in a new, dedicated connection to
Redis
to the Redis relay. This is becauseredis.psubscribe
will hog the whole Redis connection (not just Ruby process), soRelay
expects a dedicated connection for itself.
That's it, the Redis
relay service will now listen for events from external services on behalf of you, and then it'll push those events onto the internal Registry
.
It'll translate the following:
redis-cli> PUBLISH hello payload
to something understandable by your consumer, so a consumer only has to do:
on_event :hello do |payload|
puts payload
end
Notice how it behaves exactly the same as if the events did not come from Redis :)
This pattern is powerful because this means your consumers do not have to know where the events are coming from, as long as they're registed onto Registry
.
Right now, only Redis
is supported, but it is extremely easy to add a Relay
yourself since it's just a Producer
. Checkout the implementation of rocketman/relay/redis
for reference, upstream contributions for services are very welcomed too.
By default, the events emitted from your app will be stored in an in-memory Queue
, which will get processed by Rocketman threaded workers.
However this also means that if your app dies with events still in your job queue, your emitted events which are stored in-memory will be lost.
That is obviously not desirable, so that's why Rocketman ships with an option to use Redis
as your backing storage mechanism.
All you need to do is pass in a Redis
connection to Rocketman. Refer to the Configuration
section below for more information.
Here are the available options to tweak for Rocketman.
# config/initializers/rocketman.rb
Rocketman.configure do |config|
config.worker_count = 10 # defaults to 5
config.latency = 1 # defaults to 3, unit is :seconds
config.storage = Redis.new # defaults to `nil`
config.debug = true # defaults to `false`
end
Currently storage
only supports Redis
, suggestions for alternative backing mechanisms are welcomed.
debug
mode enables some debugging puts
statements, and also tweak the Thread
workers to abort_on_exception = true
. So if you have failing jobs, this is how you can figure out what's happening inside your workers.
Rocketman
, rather than a proper message bus (e.g Redis PubSub/Kafka)?It is worth noting that Rocketman
is not meant to be a replacement for the aforementioned projects -- both Redis PubSub and Kafka are battle-tested and I highly encourage to use them if you can.
But, Rocketman
recognizes that it's not an easy task to spin up an external message bus to support event-driven architecture, and that's what it's trying to do - to be a stepping stone for eventual greatness.
Moving onto a event-driven architecture is not an easy task - your team has to agree on a message bus, the DevOps team needs the capacity to manage the message bus, and then what about clustering? failovers?
So what Rocketman offers you is that you can start writing your dream-state event-driven code today, and when the time comes and your team has the capacity to move to a different message bus, then it should be a minimal change.
Here's a very crude drawing of the architecture of Rocketman
Right now events are using a fire-and-forget
mechanism, which is designed to not cause issue to producers. However, this also means that if a consumer fail to consume an event, it'll be lost forever. Next thing on the roadmap is look into a retry strategy + persistence mechanism.
Emitted events are also stored in memory in Redis support is now available!Rocketman::Pool
, which means that there's a chance that you'll lose all emitted jobs. Something to think about is to perhaps move the emitted events/job queue onto a persistent storage, like Redis for example.
The interface could also probably be better defined, as one of the goal of Rocketman is to be the stepping stone before migrating off to a real, proper message queue/pub-sub mechanism like Kafka. I want to revisit and think about how can we make that transition more seamless.
After checking out the repo, run bin/setup
to install dependencies. Then, run rake spec
to run the tests. You can also run bin/console
for an interactive prompt that will allow you to experiment.
To install this gem onto your local machine, run bundle exec rake install
. To release a new version, update the version number in version.rb
, and then run bundle exec rake release
, which will create a git tag for the version, push git commits and tags, and push the .gem
file to rubygems.org.
Bug reports and pull requests are very welcomed on GitHub at https://github.com/edisonywh/rocketman, but before a pull request is submitted, please first open up an issue for discussion.
This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the Contributor Covenant code of conduct.
The gem is available as open source under the terms of the MIT License.
Everyone interacting in the Rocketman project’s codebases, issue trackers, chat rooms and mailing lists is expected to follow the code of conduct.
Uh.. well it's named after the song by Elton John, but really, it has nothing to do with an actual Rocketman.