Skip to content

Expeditor provides asynchronous execution and fault tolerance for Microservices

License

Notifications You must be signed in to change notification settings

cookpad/expeditor

Repository files navigation

Expeditor

Gem Version Build Status

Expeditor is a Ruby library that provides asynchronous execution and fault tolerance for microservices.

It is inspired by Netflix/Hystrix.

Installation

Expeditor currently supports Ruby 2.1 and higher.

Add this line to your application's Gemfile:

gem 'expeditor'

And then execute:

$ bundle

Or install it yourself as:

$ gem install expeditor

Usage

asynchronous execution

command1 = Expeditor::Command.new do
  ...
end

command2 = Expeditor::Command.new do
  ...
end

command1.start # non blocking
command2.start # non blocking

command1.get   # wait until command1 execution is finished and get the result
command2.get   # wait until command2 execution is finished and get the result

asynchronous execution with dependencies

command1 = Expeditor::Command.new do
  ...
end

command2 = Expeditor::Command.new do
  ...
end

command3 = Expeditor::Command.new(dependencies: [command1, command2]) do |val1, val2|
  ...
end

command3.start # command1 and command2 are started concurrently, execution of command3 is wait until command1 and command2 are finished.

fallback

command = Expeditor::Command.new do
  # something that may be failed
end

# use fallback value if command is failed
command_with_fallback = command.set_fallback do |e|
  log(e)
  default_value
end

command.start.get #=> error may be raised
command_with_fallback.start.get #=> default_value if command is failed

If you set false to Expeditor::Service#fallback_enabled, fallbacks do not occur. It is useful in test codes.

timeout

command = Expeditor::Command.new(timeout: 1) do
  ...
end

command.start
command.get #=> Timeout::Error is raised if execution is timed out

retry

command = Expeditor::Command.new do
  ...
end

# the option is completely same as retryable gem
command.start_with_retry(
  tries: 3,
  sleep: 1,
  on: [StandardError],
)

using thread pool

Expeditor use concurrent-ruby's executors as thread pool.

require 'concurrent'

service = Expeditor::Service.new(
  executor: Concurrent::ThreadPoolExecutor.new(
    min_threads: 0,
    max_threads: 5,
    max_queue: 100,
  )
)

command = Expeditor::Command.new(service: service) do
  ...
end

service.status
# => #<Expeditor::Status:0x007fdeeeb18468 @break=0, @dependency=0, @failure=0, @rejection=0, @success=0, @timeout=0>

service.reset_status!  # reset status in the service

circuit breaker

The circuit breaker needs a service metrics (success, failure, timeout, ...) to decide open the circuit or not. Expeditor's circuit breaker has a few configuration for how it collects service metrics and how it opens the circuit.

For service metrics, Expeditor collects them with the given time window. The metrics is gradually collected by breaking given time window into some peice of short time windows and resetting previous metrics when passing each short time window.

service = Expeditor::Service.new(
  threshold: 0.5,      # If the failure rate is more than or equal to threshold, the circuit will be opened.
  sleep: 1,            # If once the circuit is opened, the circuit is still open until sleep time seconds is passed even though failure rate is less than threshold.
  non_break_count: 20, # If the total count of metrics is not more than non_break_count, the circuit is not opened even though failure rate is more than threshold.
  period: 10,          # Time window of collecting metrics (in seconds).
)

command = Expeditor::Command.new(service: service) do
  ...
end

non_break_count is used to ignore requests to the service which is not frequentlly requested. Configure this value considering your estimated "requests per period to the service". For example, when period = 10 and non_break_count = 20 and the requests do not occur more than 20 per 10 seconds, the circuit never opens because Expeditor ignores that "small number of requests". If you don't ignore the failures in that case, set non_break_count to smaller value than 20.

The default values are:

  • threshold: 0.5
  • sleep: 1
  • non_break_count: 20
  • period: 10

synchronous execution

Use current_thread option of #start, command executes synchronous on current thread.

command1 = Expeditor::Command.new do
  ...
end

command2 = Expeditor::Command.new do
  ...
end

command1.start(current_thread: true) # blocking
command2.start(current_thread: true) # blocking

command1.get
command2.get

Development

After checking out the repo, run bin/setup to install dependencies. Then, 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 to create a git tag for the version, push git commits and tags, and push the .gem file to rubygems.org.

Contributing

  1. Fork it ( https://github.com/cookpad/expeditor/fork )
  2. Create your feature branch (git checkout -b my-new-feature)
  3. Commit your changes (git commit -am 'Add some feature')
  4. Push to the branch (git push origin my-new-feature)
  5. Create a new Pull Request

About

Expeditor provides asynchronous execution and fault tolerance for Microservices

Resources

License

Stars

Watchers

Forks

Packages

No packages published