Create Deploy Pipeline for AWS Elastic Beanstalk App

AWS Elastic Beanstalk app can deploy simply by using CodePipeline.

Below is a pipeline I created on CodePipeline for deploying apps to Elastic Beanstalk, S3, and Cloud. I just introduce CodePipeline this time.

A deploy pipeline

This project’s directory structure is like this:

my-api
  + .elasticbeanstalk
  | + config.yaml
  + application.py
  + requirements.txt

application.py and requrements.txt are simple Flask app. You can also read a Flask tutorial.

Before creating the deploy pipeline, you need to create your pipeline and Beanstalk app. Please refer to a creating pipeline tutorial and Beanstalk tutorial.

After you create a pipeline and minimal Beanstalk app, you need to set up a source where you get your code from. Below is the example using GitHub:

Source configuration for getting a source from GitHub master branch

Next, you need to add a build step to do unit testing or build your app into binary code or archive. And also I specify Output artifacts as “api-artifact”. This is an archived file built in this step to use after steps. In this pipeline, a next deploy step will use this file. Below is my setting:

Build configuration for unit testing

You need to create buildspec.yaml to run the build step on CodeBuild.

	version: 0.2
	phases:
	pre_build:
	commands:
	- pip install -r my-api/requirements.txt
	build:
	commands:
	- python -m unittest discover -s my-api
	artifacts:
	files:
	- '**/*'
	base-directory: my-api

view raw buildspec_python.yaml hosted with ❤ by GitHub

The app is written with Python and include unit tests in a build step, and so add python -m unittest discover -s my-api under the build . A deploy step uses app artifact in the my-api directory, and so you need to base-directory: my-api to create an artifact under the my-api directory.

Next, since we want to deploy the app to Elastic Beanstalk. Below is a deploy step configuration:

Deploy configuration for deploying new code to Elastic Beanstalk

So you add a configuration file for Elastic Beans like below:

	# You must set this file into .elasticbeanstalk direcotry like this;
	# .elasticbeanstalk/config.yaml
	branch-defaults:
	default:
	environment: api-env
	environment-defaults:
	api-env:
	branch: null
	repository: null
	global:
	application_name: my-api
	default_ec2_keyname: null
	default_platform: python-3.6
	default_region: ap-northeast-1
	include_git_submodules: true
	instance_profile: null
	platform_name: null
	platform_version: null
	profile: eb-cli
	sc: null
	workspace_type: Application

view raw config.yaml hosted with ❤ by GitHub

Then you just push your code to GitHub’s master branch, CodePipeline automatically builds it and deploy to Beanstalk.