aws-lambda-jruby

Run ruby code via JRuby on AWS Lambda (java)

This is a Dockerized gradle project for packaging a Ruby project into a zip file for execution in AWS Lambda using the Java 8 runtime.

Building

• Create a ruby project in a directory my_ruby_lambda
  • The project directory should contain a Gemfile for any external libraries you are using
    • This file may be blank
  • The project directory should contain a main.rb ruby file which will be the entrypoint for the lambda
    • The main file may look something like this:

BASE_DIRECTORY = File.realpath(File.dirname(__FILE__)).to_s

# Add current directory to the path
$LOAD_PATH.unshift BASE_DIRECTORY

# Require everything in lib
Dir["#{BASE_DIRECTORY}/lib/**/*.rb"].sort.each { |file| require file.sub("#{BASE_DIRECTORY}/", '').chomp('.rb') }

# rubocop:disable Style/GlobalVars
def main
  MyRubyLambda::Lambda.new($lambda_arg).perform
rescue => e
  puts "Lambda execution failed: Error #{e.inspect}\n#{e.backtrace.join("\n")}"
  raise
end
# rubocop:enable Style/GlobalVars

# Only run the main function if the file has been called directly
main if $PROGRAM_NAME == __FILE__

• Add any other ruby code you need
  • The directory structure might look something like this:

|-- my_ruby_lambda
    |-- Gemfile
    |-- main.rb
    |-- lib
        |-- awesome_sauce.rb
        |-- ...

• Build app.zip in the project directory
  • docker run -v my_ruby_lambda:/usr/src/code firespring/aws-lambda-jruby

Execution

• Create an AWS Lambda fuction on AWS

  • Use an appropriate name
  • Set the runtime to Java 8
  • Set an appropriate role for the lambda

• Upload the lambda function code

  • Code entry type should be .ZIP file
  • Runtime should be Java 8
  • Set the handler value to AWSLambdaJRuby::handler
  • Click Upload and select app.zip
  • Save your changes

• You can now run your lambda

  • NOTE: The first time your lambda runs it may take up to 60 seconds to execute
  • (subsequent executions should be sufficiently fast)