Join the chat at

Circle CI Test Coverage Code Climate Gem Version

Why Gantree?

The name is derived from the word gantry which is a large crane used in ports to pick up shipping containers and load them on a ship. Gantry was already taken so I spelled it "tree" because the primary use is for elastic beanstalk and I guess a beanstalk is a form of tree?


This tool is intended to help you setup a which allows you to deploy a prebuilt image of your application to Elastic Beanstalk. This also allows you to do versioned deploys to your Elastic Beanstalk application and create an archive of every versioned in amazons s3 bucket service.



You need to have your AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables set in order to use the tool as well as the proper aws permissions for Elastic Beanstalk, and S3 access.

Note : For gantree versions >= 0.6.14, configuration of AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables is not necessary if running on an instance with IAM Roles enabled.

To check if your EC2 has iam role or not, run below command.


Install docker for MAC OSX

Generate your login credentials token:



gem install gantree

Initialize a new Repo

What this does is create a new inside your repository and uploads your docker login credentials to s3 (for private repo access) so you can do deploys. We need the -u to specify a username to rename your .dockercfg and reference it in the

# the username here is your docker.hub login
gantree init -u frodriguez -p 3000 bleacher/cauldron:master
# this will upload your docker config files to a bucket called "frodrigeuz-docker-cgfs"

If you don't have a docker.hub account, you can still use gantree without the -u flag, but you will have to explicitly specify the bucket for S3 storage since the default S3 bucket name is generated from the docker.hub login.

Specify the bucket for to S3 store docker configuration
# Since S3 bucket names are globally namespaced, the default bucket may be taken and unavailable
# Gantree gives you the option to specify an S3 bucket name of your choice
gantree init -u frodgriguez -p 3000 -b hopefully_this_bucket_name_is_available bleacher/cauldron:master


This command renames your temporarily to, uploads it to a NAME_OF_APP-versions bucket, creates a new elastic beanstalk version, and asks the specified elastic beanstalk environment to update to that new version.

gantree deploy cauldron-stag-s1

By default this will check for the environment cauldron-stag-s1 and deploy to the app stag-cauldron-app. You can also specify an environment name directly using -e.

gantree deploy -e cauldron-stag-s1 stag-cauldron-app-s1

You can also specify a new image tag to use for the deploy

gantree deploy -t latest stag-cauldon-app-s1


This command does three things:

1) Builds the image (with an auto generated tag (origin-branch-hash) or custom tag using the -t flag 2) Uploads the image based on the --image-path flag or gantree.cfg (ie. hub.docker, 3) Deploys the image to elastic beanstalk based on command line arguments and cfn files.

# assuming you already ran 'gantree init' and generated the file
gantree ship linguist-stag-s2 -t branch-with-features-and-bug-fixes
Autodetecting environment roles

Let's say your linguist app consists of both an app and a worker. The Gantree ship command has an autodetect-app-role flag which will automatically detect the environment's role (app, worker, listener, etc.) by matching the environment name's third dash-delimited string and inject that as a $ROLE variable inside the Docker containers.

For example, suppose this is the cfn.json file:

  "ApplicationName": "linguist-stag-s2",
  "EnvironmentName": "stag-linguist-app-s2",

  "ApplicationName": "linguist-stag-s2",
  "EnvironmentName": "stag-linguist-worker-s2",

You can call gantree ship with the autodetect flag

gantree ship linguist-stag-s2 --autodetect-app-role=true -t branch-with-features-and-bug-fixes

Now every environment's Docker container will have the $ROLE variable:

# Inside stag-linguist-app-s2 Docker container
[email protected]$ echo $ROLE

# Inside stag-linguist-worker-s2 Docker container
[email protected]$ echo $ROLE

As of Gantree version 0.4.9, the autodetect flag default is always on but if your env naming convention doesnt follow the stag-linguist-app-s2 convention, you should turn this off to avoid setting the wrong role variable in the Docker container.

Update and Create

As mentioned earlier, Gantree leverages the power of aws cloud formation. The gantree commands update and create are cloud formation specific gantree commands.


We use gantree's update command to change an application name, environment name, or other cloud formation changes.

# after making changes in your /cfn/*cfn.json files, you want to update the stack
gantree update linguist-stag-s1
# You can use the -r flag to easily add a new role to the cnf template
# Note the dry-run flag is used so we actually don't update the stack, but just changing the cfn files locally
gantree update linguist-prod-s1 -r worker --dry-run


The gantree create command does three things:

1) Creates cfn templates 2) Sends them to s3 3) Creates the stack

gantree create linguist-prod-s1

# you can use the --local flag to skip the first step
gantree create linguist-prod-s1 --local

We use gantree's create command to create a new cfn template from an existing cfn template

# we already have linguist-prod-s1, and we want to create linguist-prod-s2
gantree create linguist-prod-s2 --dupe linguist-prod-s1


Allow defaults for commands to be set in a json file

  "ebextensions" : "[email protected]:br/.ebextensions.git",
  "default_instance_size" : "m3.medium"

.ebextension Support

Elastic Beanstalk cli allows you to create a .ebextension folder that you can package with your deploy to control the host/environment of your application. Deploying only a docker container image referenced in has the unfortunate side effect of losing this extreamly powerful feature. To allow this feature to be included in gantree and make it even better you can select either to package a local .ebextension folder with your deploy, package a remote .ebextension folder hosted in github (with branch support) or even create a .gantreecfg file to make either of these type of deploys a default.

gantree deploy -x "git:br/ebextensions:master" stag-cauldron-app-s1

By default your application will be created on a t1.micro unless you specify otherwise:

gantree ceate your_app_name -i m3.medium

What if you need a database? Here enters the beauty of RDS

PostgreSQL: gantree create your_app_name --rds pg

Mysql: gantree create your_app_name --rds msql

Traveling Ruby :

You can now compile this into a tarball and distribute.

To generate tarball & push to s3.ß

# Change below variable in Makefile before you run below command
make clean && make linux

To use gantree binary

aws s3 cp s3://<your_bucket>/gantree-0.6.15-linux-x86_64.tar.gz .
tar -xf gantree-0.6.15-linux-x86_64.tar.gz -C /opt
ln -s /opt/gantree-0.6.15-linux-x86_64/gantree /usr/bin/gantree


What if you want a cdn behind each of your generated applications

Fastly: gantree create your_app_name --cdn fastly

CloudFront: gantree create yoour_app_name --cdn cloudfront

Redis & Memcached

Elasticache gantree create your_app_name --cache redis or gantree create your_app_name --cache memcache

Autogenerated Release Notes

I would like have built in integration with opbeat configurable thorugh the .gantreecfg located in the applications repository.


Also if the cloud formation template that is generated doesn't match your needs (which it might now) you can edit the .rb files in the repo's cfn folder, build your own gem and use it how you like.