Getting started with the Ruby amqp gem

This Documentation Has Moved to rubyamqp.info

amqp gem documentation guides are now hosted on rubyamqp.info.

About this guide

This guide is a quick tutorial that helps you to get started with v0.9.1 of the AMQP specification in general and the Ruby amqp gem in particular. It should take about 20 minutes to read and study the provided code examples. This guide covers:

  • Installing RabbitMQ, a mature popular server implementation of the AMQP protocol.
  • Installing the amqp gem via Rubygems and Bundler.
  • Running a “Hello, world” messaging example that is a simple demonstration of 1:1 communication.
  • Creating a “Twitter-like” publish/subscribe example with 1 publisher and 4 subscribers that demonstrates 1:n communication.
  • Creating a topic routing example with 2 publishers and 8 subscribers showcasing n:m communication when subscribers only receive messages that they are interested in.
  • Learning how the amqp gem can be integrated with Ruby objects in a way that makes unit testing easy.

This work is licensed under a Creative Commons Attribution 3.0 Unported License (including images & stylesheets). The source is available on Github.

Which versions of the amqp gem does this guide cover?

This guide covers v0.8.0 and later of the Ruby amqp gem.

Installing RabbitMQ

The RabbitMQ site has a good installation guide that addresses many operating systems. On Mac OS X, the fastest way to install RabbitMQ is with Homebrew:


brew install rabbitmq

then run it:


rabbitmq-server

On Debian and Ubuntu, you can either download the RabbitMQ .deb package and install it with dpkg or make use of the apt repository that the RabbitMQ team provides. For RPM-based distributions like RedHat or CentOS, the RabbitMQ team provides an RPM package.

The RabbitMQ package that ships with recent Ubuntu versions (for example, 10.10) is outdated and will not work with v0.8.0 and later of the amqp gem (we need at least RabbitMQ v2.0 for use with this guide).

Installing the Ruby amqp gem

Make sure that you have Ruby and Rubygems installed

This guide assumes that you have installed one of the following supported Ruby implementations:

  • Ruby v1.8.7 [except for 1.8.7-p248 and -p249 that have a bug that severely affects amqp gem]
  • Ruby v1.9.2
  • Ruby v1.9.3
  • JRuby (we recommend v1.6)
  • Rubinius v1.2 or higher
  • Ruby Enterprise Edition

You can use Rubygems to install the amqp gem

On Microsoft Windows 7:


gem install eventmachine --pre
gem install amqp

On other OSes or JRuby:


gem install amqp

You can also use Bundler to install the gem


source :rubygems

gem "amqp", "~> 0.9.0" # optionally: :git => "git://github.com/ruby-amqp/amqp.git", :branch => "0.9.x-stable"

Verifying your installation

Let us verify your installation with this quick irb session:


irb -rubygems

:001 > require "amqp"
=> true
:002 > AMQP::VERSION
=> "0.9.0"

“Hello, world” example

Let us begin with the classic “Hello, world” example. First, here is the code:

(if the example above isn’t displayed, see this gist)

This example demonstrates a very common communication scenario: application A wants to publish a message that will end up in a queue that application B listens on. In this case, the queue name is “amqpgem.examples.hello”. Let us go through the code step by step:


require "rubygems"
require "amqp"

is the simplest way to load the amqp gem if you have installed it with RubyGems. The following piece of code


EventMachine.run do
  # ...
end

runs what is called the EventMachine reactor. We will not go into what the term ‘reactor’ means here, but suffice it to say that the amqp gem is asynchronous and is based on an asynchronous network I/O library called EventMachine.

The next line


connection = AMQP.connect(:host => '127.0.0.1')

connects to the server running on localhost, with the default port (5672), username (guest), password (guest) and virtual host (‘/’).

The next line


channel  = AMQP::Channel.new(connection)

opens a new channel. AMQP is a multi-channeled protocol that uses channels to multiplex a TCP connection. Channels are opened on a connection, therefore the AMQP::Channel constructor takes a connection object as a parameter.

This line


queue    = channel.queue("amqpgem.examples.helloworld", :auto_delete => true)

declares a queue on the channel that we have just opened. Consumer applications get messages from queues. We declared this queue with the “auto-delete” parameter. Basically, this means that the queue will be deleted when there are no more processes consuming messages from it.

The next line


exchange = channel.direct("")

instantiates an exchange. Exchanges receive messages that are sent by producers. Exchanges route messages to queues according to rules called bindings. In this particular example, there are no explicitly defined bindings. The exchange that we defined is known as the default exchange and it has implied bindings to all queues. Before we get into that, let us see how we define a handler for incoming messages


queue.subscribe do |payload|
  puts "Received a message: #{payload}. Disconnecting..."
  connection.close { EventMachine.stop }
end

AMQP::Queue#subscribe takes a block that will be called every time a message arrives. AMQP::Session#close closes the AMQP connection and runs a callback that stops the EventMachine reactor.

Finally, we publish our message


exchange.publish "Hello, world!", :routing_key => queue.name

Routing key is one of the message attributes. The default exchange will route the message to a queue that has the same name as the message’s routing key. This is how our message ends up in the “amqpgem.examples.helloworld” queue.

This first example can be modified to use the method chaining technique:

This diagram demonstrates the “Hello, world” example data flow:

For the sake of simplicity, both the message producer (App I) and the consumer (App II) are running in the same Ruby process. Now let us move on to a little bit more sophisticated example.

Blabbr: one-to-many publish/subscribe (pubsub) example

The previous example demonstrated how a connection to a broker is made and how to do 1:1 communication using the default exchange. Now let us take a look at another common scenario: broadcast, or multiple consumers and one producer.

A very well-known broadcast example is Twitter: every time a person tweets, followers receive a notification. Blabbr, our imaginary information network, models this scenario: every network member has a separate queue and publishes blabs to a separate exchange. 3 Blabbr members, Joe, Aaron and Bob, follow the official NBA account on Blabbr to get updates about what is happening in the world of basketball. Here is the code:

The first line has a few differences from the “Hello, world” example above:

  • We use AMQP.start instead of AMQP.connect
  • Instead of return values, we pass a block to the connection method and it yields a connection object back as soon as the connection is established.
  • Instead of passing connection parameters as a hash, we use a URI string.

AMQP.start is just a convenient way to do


EventMachine.run do
  AMQP.connect(options) do |connection|
    # ...
  end
end

The AMQP.start call blocks the current thread which means that its use is limited to scripts and small command line applications. Blabbr is just that.

AMQP.connect, when invoked with a block, will yield a connection object as soon as the AMQP connection is open. Finally, connection parameters may be supplied as a Hash or as a connection string. The AMQP.connect method documentation contains all of the details.

In this example, opening a channel is no different to opening a channel in the previous example, however, the exchange is declared differently


exchange = channel.fanout("nba.scores")

The exchange that we declare above using AMQP::Channel#fanout is a fanout exchange. A fanout exchange delivers messages to all of the queues that are bound to it: exactly what we want in the case of Blabbr.

This piece of code


channel.queue("joe", :auto_delete => true).bind(exchange).subscribe do |payload|
  puts "#{payload} => joe"
end

is similar to the subscription code that we used for message delivery previously, but what does that AMQP::Queue#bind method do? It sets up a binding between the queue and the exchange that you pass to it. We need to do this to make sure that our fanout exchange routes messages to the queues of any subscribed followers.


exchange.publish("BOS 101, NYK 89").publish("ORL 85, ALT 88")

demonstrates AMQP::Exchange#publish call chaining. Blabbr members use a fanout exchange for publishing, so there is no need to specify a message routing key because every queue that is bound to the exchange will get its own copy of all messages, regardless of the queue name and routing key used.

A diagram for Blabbr looks like this:

Next we use EventMachine’s add_timer method to run a piece of code in 2 seconds from now:


EventMachine.add_timer(2) do
  exchange.delete

  connection.close { EventMachine.stop }
end

The code that we want to run deletes the exchange that we declared earlier using AMQP::Exchange#delete and closes the AMQP connection with AMQP::Session#close. Finally, we stop the EventMachine event loop and exit.

Blabbr is pretty unlikely to secure hundreds of millions of dollars in funding, but it does a pretty good job of demonstrating how one can use AMQP fanout exchanges to do broadcasting.

Weathr: many-to-many topic routing example

So far, we have seen point-to-point communication and broadcasting. Those two communication styles are possible with many protocols, for instance, HTTP handles these scenarios just fine. You may ask “what differentiates AMQP?” Well, next we are going to introduce you to topic exchanges and routing with patterns, one of the features that makes AMQP very powerful.

Our third example involves weather condition updates. What makes it different from the previous two examples is that not all of the consumers are interested in all of the messages. People who live in Portland usually do not care about the weather in Hong Kong (unless they are visiting soon). They are much more interested in weather conditions around Portland, possibly all of Oregon and sometimes a few neighbouring states.

Our example features multiple consumer applications monitoring updates for different regions. Some are interested in updates for a specific city, others for a specific state and so on, all the way up to continents. Updates may overlap so that an update for San Diego, CA appears as an update for California, but also should show up on the North America updates list.

Here is the code:

The first line that is different from the Blabbr example is


exchange = channel.topic("pub/sub", :auto_delete => true)

We use a topic exchange here. Topic exchanges are used for multicast messaging where consumers indicate which topics they are interested in (think of it as subscribing to a feed for an individual tag in your favourite blog as opposed to the full feed). Routing with a topic exchange is done by specifying a routing pattern on binding, for example:


channel.queue("americas.south").bind(exchange, :routing_key => "americas.south.#").subscribe do |headers, payload|
  puts "An update for South America: #{payload}, routing key is #{headers.routing_key}"
end

Here we bind a queue with the name of “americas.south” to the topic exchange declared earlier using the AMQP::Queue#bind method. This means that only messages with a routing key matching “americas.south.#” will be routed to that queue. A routing pattern consists of several words separated by dots, in a similar way to URI path segments joined by slashes. Here are a few examples:

  • asia.southeast.thailand.bangkok
  • sports.basketball
  • usa.nasdaq.aapl
  • tasks.search.indexing.accounts

Now let us take a look at a few routing keys that match the “americas.south.#” pattern:

  • americas.south
  • americas.south.brazil
  • americas.south.brazil.saopaolo
  • americas.south.chile.santiago

In other words, the “#” part of the pattern matches 0 or more words.

For a pattern like “americas.south.*”, some matching routing keys would be:

  • americas.south.brazil
  • americas.south.chile
  • americas.south.peru

but not

  • americas.south
  • americas.south.chile.santiago

so “*” only matches a single word. The AMQP v0.9.1 specification says that topic segments (words) may contain the letters A-Z and a-z and digits 0-9.

A (very simplistic) diagram to demonstrate topic exchange in action:

One more thing that is different from previous examples is that the block we pass to AMQP::Queue#subscribe now takes two arguments: a header and a body (often called the payload). Long story short, the header parameter lets you access metadata associated with the message. Some examples of message metadata attributes are:

  • message content type
  • message content encoding
  • message priority
  • message expiration time
  • message identifier
  • reply to (specifies which message this is a reply to)
  • application id (identifier of the application that produced the message)

and so on.

As the following binding demonstrates, “#” and “*” can also appear at the beginning of routing patterns:


channel.queue("us.tx.austin").bind(exchange, :routing_key => "#.tx.austin").subscribe do |headers, payload|
  puts "An update for Austin, TX: #{payload}, routing key is #{headers.routing_key}"
end

For this example the publishing of messages is no different from that of previous examples. If we were to run the program, a message published with a routing key of “americas.north.us.ca.berkeley” would be routed to 2 queues: “us.california” and the server-named queue that we declared by passing a blank string as the name:


channel.queue("", :exclusive => true) do |queue|
  queue.bind(exchange, :routing_key => "americas.north.#").subscribe do |headers, payload|
    puts "An update for North America: #{payload}, routing key is #{headers.routing_key}"
  end
end

The name of the server-named queue is generated by the broker and sent back to the client with a queue declaration confirmation. Because the queue name is not known before the reply arrives, we pass AMQP::Channel#queue a callback and it yields us back a queue object once confirmation has arrived.

Avoid race conditions

A word of warning: you may find examples on the Web of AMQP::Channel#queue usage that do not use callbacks. We recommend that you use a callback for server-named queues, otherwise your code may be subject to race conditions. Even though the amqp gem tries to be reasonably smart and protect you from most common problems (for example, binding operations will be delayed until after queue name is received from the broker), there is no way it can do so for every case. The primary reason for supporting AMQP::Channel#queue usage without a callback for server-named queues is backwards compatibility with earlier versions.

Integration with objects

Since Ruby is a genuine object-oriented language, it is important to demonstrate how the Ruby amqp gem can be integrated into rich object-oriented code.

The AMQP::Queue#subscribe callback does not have to be a block. It can be any Ruby object that responds to `call`. A common technique is to combine Object#method and Method#to_proc and use object methods as message handlers.

An example to demonstrate this technique:


class Consumer

  #
  # API
  #

  def handle_message(metadata, payload)
    puts "Received a message: #{payload}, content_type = #{metadata.content_type}"
  end # handle_message(metadata, payload)
end


class Worker

  #
  # API
  #


  def initialize(channel, queue_name = AMQ::Protocol::EMPTY_STRING, consumer = Consumer.new)
    @queue_name = queue_name

    @channel    = channel
    @channel.on_error(&method(:handle_channel_exception))

    @consumer   = consumer
  end # initialize

  def start
    @queue = @channel.queue(@queue_name, :exclusive => true)
    @queue.subscribe(&@consumer.method(:handle_message))
  end # start



  #
  # Implementation
  #

  def handle_channel_exception(channel, channel_close)
    puts "Oops... a channel-level exception: code = #{channel_close.reply_code}, message = #{channel_close.reply_text}"
  end # handle_channel_exception(channel, channel_close)
end

The “Hello, world” example can be ported to use this technique:

The most important line in this example is


@queue.subscribe(&@consumer.method(:handle_message))

Ampersand (&) preceding an object is equivalent to calling the #to_proc method on it. We obtain a Consumer#handle_message method reference with


@consumer.method(:handle_message)

and then the ampersand calls #to_proc on it. AMQP::Queue#subscribe then will be using this Proc instance to handle incoming messages.

Note that the `Consumer` class above can be easily tested in isolation, without spinning up any AMQP connections. Here is one example using RSpec


require "ostruct"
require "json"

# RSpec example
describe Consumer do
  describe "when a new message arrives" do
    subject { described_class.new }

    let(:metadata) do
      o = OpenStruct.new

      o.content_type = "application/json"
      o
    end
    let(:payload)  { JSON.encode({ :command => "reload_config" }) }

    it "does some useful work" do
      # check preconditions here if necessary

      subject.handle_message(metadata, payload)

      # add your code expectations here
    end
  end
end

Wrapping up

This is the end of the tutorial. Congratulations! You have learned quite a bit about both AMQP v0.9.1 and the amqp gem. This is only the tip of the iceberg. AMQP has many more features built into the protocol:

  • Reliable delivery of messages
  • Message confirmations (a way to tell broker that a message was or was not processed successfully)
  • Message redelivery when consumer applications fail or crash
  • Load balancing of messages between multiple consumers
  • Message metadata attributes

and so on. Other guides explain these features in depth, as well as use cases for them. To stay up to date with amqp gem development, follow @rubyamqp on Twitter and join our mailing list.

What to read next

Documentation is organized as a number of documentation guides, covering all kinds of topics from use cases for various exchange types to error handling and Broker-specific AMQP 0.9.1 extensions.

We recommend that you read the following guides next, if possible, in this order:

  • AMQP 0.9.1 Model Explained. A simple 2 page long introduction to the AMQP Model concepts and features. Understanding the AMQP Model will make a lot of other documentation, both for the Ruby amqp gem and RabbitMQ itself, easier to follow. With this guide, you don’t have to waste hours of time reading the whole specification.
  • Connection to the broker. This guide explains how to connect to an AMQP broker and how to integrate the amqp gem into standalone and Web applications.
  • Working With Queues. This guide focuses on features that consumer applications use heavily.
  • Working With Exchanges. This guide focuses on features that producer applications use heavily.
  • Patterns & Use Cases. This guide focuses implementation of common messaging patterns using AMQP Model features as building blocks.
  • Error Handling & Recovery. This guide explains how to handle protocol errors, network failures and other things that may go wrong in real world projects.

If you are migrating your application from earlier versions of the amqp gem (0.6.x and 0.7.x), to 0.8.x and later, there is the amqp gem 0.8 migration guide.

Authors

This guide was written by Michael Klishin and edited by Chris Duncan.

Tell us what you think!

Please take a moment to tell us what you think about this guide on Twitter or the Ruby AMQP mailing list. Let us know what was unclear or what has not been covered. Maybe you do not like the guide style or grammar or discover spelling mistakes. Reader feedback is key to making the documentation better.

If, for some reason, you cannot use the communication channels mentioned above, you can contact the author of the guides directly