ElasticBeanstalk

Configure and deploy a rails app to Elastic Beanstalk via rake in 60 seconds. Maintain multiple environment DRY configurations and .ebextensions in one easy to use configuration file.

This gem simplifies configuration, and passes the heavy lifting to the eb_deployer from ThoughtWorksStudios.

Installation

Add this line to your application's Gemfile:

gem 'elastic-beanstalk'

And then execute:

$ bundle

Or install it yourself as:

$ gem install elastic-beanstalk

Features

The elastic-beanstalk gem provides:

  • Rake tasks to simplify all interactions
  • Multi-environment configuration inheritance so your configs can be DRY
  • Full access to the range of configuration options provided by AWS Elastic Beanstalk
  • Keep all configurations including .ebextensions in one file (they are inheritable and can be overridden too)
  • Provide access to helpers such as the SmokeTester to simplify configurations

Plus

Since eb_deployer is doing the heavy lifting, by proxy you get access to great continuous delivery features such as:

  • Blue Green deployment strategy
  • In Place deployment strategy
  • Smoke Testing upon deployment before Blue Green DNS switching

Usage

Given an application named 'acme':

Step 1: Add a ~/.aws/acme.yml

This should contain the access and secret keys generated from the selected IAM user. This is the only file that will need to reside outside the repository.

access_key_id: XXXXXX
secret_access_key: XXXXXX

Step 2. Add a config/eb.yml to your rails project

Something like this should get you started

app: acme
region: us-east-1
solution_stack_name: 64bit Amazon Linux running Ruby 1.9.3

development:
  strategy: inplace_update
  options:
    aws:autoscaling:launchconfiguration:
      InstanceType: t1.micro

production:
  options:
    aws:autoscaling:launchconfiguration:
      InstanceType: t1.small

Step 3. Package and deploy

The default is the 'development' environment, change this via command line by prefixing with i.e. RAILS_ENV=production

$ rake eb:package eb:deploy

Step 4. Get some coffee

This will take a while. We intend to provide an example in the wiki and/or samples dir that implements a caching strategy detailed here to speed up deployment.

EB Rake Tasks

rake eb:clobber                 # Remove any generated package
rake eb:config                  # Setup AWS.config and merge/override environments into one resolved configuration
rake eb:deploy[version]         # Deploy to Elastic Beanstalk
rake eb:destroy[force]          # ** Warning: Destroy Elastic Beanstalk application and *all* environments
rake eb:package                 # Package zip source bundle for Elastic Beanstalk
rake eb:show_config[version]    # Show resolved configuration without doing anything

EB:RDS Rake Tasks

The EB:RDS tasks are intended to make use of RDS tasks simple given existing configuration in the eb.yml. i.e. create a snapshot before or after an eb:deploy. The following rake tasks exist:

rake eb:rds:create_snapshot[instance_id,snapshot_id]   # Creates an RDS snapshot
rake eb:rds:instances                                  # List RDS instances
rake eb:rds:snapshots                                  # List RDS snapshots

For example, this would create a snapshot prior to the deployment (and migration) to version 1.1.0:

rake eb:rds:create_snapshot[acme, pre-1.1.0] eb:deploy[1.1.0]

A real-world example

Deploy version 1.1.3 of acme to production

$ RAILS_ENV=production rake eb:package eb:deploy[1.1.3]

config/eb.yml

# This is a sample that has not been executed so it may not be exactly 100%, but is intended to show
#   that access to full options_settings and .ebextensions is intended.
#---
app: acme
region: us-east-1
solution_stack_name: 64bit Amazon Linux running Ruby 1.9.3
package:
  verbose: true
  exclude_dirs: [solr, features] # additional dirs that merge with default excludes
  exclude_files: [rspec.xml, README*, db/*.sqlite3]
smoke_test: |
    lambda { |host|

      EbSmokeTester.test_url("http://#{host}/ping", 600, 5, 'All good! Everything is up and checks out.')
    }
#--
ebextensions:
  01settings.config:
    # Run rake tasks before an application deployment
    container_commands:
      01seed:
        command: rake db:seed
        leader_only: true
  # run any necessary commands
  02commands.config:
    container_commands:
      01timezone:
        command: "ln -sf /usr/share/zoneinfo/America/Chicago /etc/localtime"

  # These are things that make sense for any Ruby application see:
  #     https://github.com/gkop/elastic-beanstalk-ruby/blob/master/.ebextensions/ruby.config
  03-ruby.config:

    # Install git in order to be able to bundle gems from git
    packages:
      yum:
        git: []
    commands:
      # Run rake with bundle exec to be sure you get the right version
      add_bundle_exec:
        test: test ! -f /opt/elasticbeanstalk/support/.post-provisioning-complete
        cwd: /opt/elasticbeanstalk/hooks/appdeploy/pre
        command: perl -pi -e 's/(rake)/bundle exec $1/' 11_asset_compilation.sh 12_db_migration.sh
      # Bundle with --deployment as recommended by bundler docs
      #   cf. http://gembundler.com/v1.2/rationale.html under Deploying Your Application
      add_deployment_flag:
        test: test ! -f /opt/elasticbeanstalk/support/.post-provisioning-complete
        cwd: /opt/elasticbeanstalk/hooks/appdeploy/pre
        command: perl -pi -e 's/(bundle install)/$1 --deployment/' 10_bundle_install.sh
      # Vendor gems to a persistent directory for speedy subsequent bundling
      make_vendor_bundle_dir:
        test: test ! -f /opt/elasticbeanstalk/support/.post-provisioning-complete
        command: mkdir /var/app/support/vendor_bundle
      # Store the location of vendored gems in a handy env var
      set_vendor_bundle_var:
        test: test ! -f /opt/elasticbeanstalk/support/.post-provisioning-complete
        cwd: /opt/elasticbeanstalk/support
        command: sed -i '12iexport EB_CONFIG_APP_VENDOR_BUNDLE=$EB_CONFIG_APP_SUPPORT/vendor_bundle' envvars
      # The --deployment flag tells bundler to install gems to vendor/bundle/, so
      # symlink that to the persistent directory
      symlink_vendor_bundle:
        test: test ! -f /opt/elasticbeanstalk/support/.post-provisioning-complete
        cwd: /opt/elasticbeanstalk/hooks/appdeploy/pre
        command: sed -i '6iln -s $EB_CONFIG_APP_VENDOR_BUNDLE ./vendor/bundle' 10_bundle_install.sh
      # Don't run the above commands again on this instance
      #   cf. http://stackoverflow.com/a/16846429/283398
      z_write_post_provisioning_complete_file:
        cwd: /opt/elasticbeanstalk/support
        command: touch .post-provisioning-complete
#---
options:
  aws:autoscaling:launchconfiguration:
    EC2KeyName: eb-ssh
    SecurityGroups: 'acme-production-control'

  aws:autoscaling:asg:
    MinSize: 1
    MaxSize: 5

  aws:elb:loadbalancer:
    SSLCertificateId: 'arn:aws:iam::XXXXXXX:server-certificate/acme'
    LoadBalancerHTTPSPort: 443

  aws:elb:policies:
    Stickiness Policy: true

  aws:elasticbeanstalk:sns:topics:
    Notification Endpoint: '[email protected]'

  aws:elasticbeanstalk:application:
    Application Healthcheck URL: '/'
#---
development:
  strategy: inplace_update
  options:
    aws:autoscaling:launchconfiguration:
      InstanceType: t1.micro
    aws:elasticbeanstalk:application:environment:
      RAILS_SKIP_ASSET_COMPILATION: true
#---
production:
  options:
    aws:autoscaling:launchconfiguration:
      InstanceType: t1.small

Still to come

  1. Caching sample config
  2. More thorough access to the Elastic Beanstalk api as-needed.

Contributing

Please contribute! While this is working great, a greater scope of functionality is certainly easily attainable with this foundation in place.

  1. Fork it
  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 new Pull Request