Udacity AWS DevOps Engineer Capstone Project

Project Overview

This capstone project showcases the use of several CI/CD tools and cloud services covered in the program Udacity - AWS Cloud DevOps Engineer.

Introduction

This project "operationalize" a sample python/flask demo app "hello", using CircleCI and a Kubernetes(K8S) cluster deployed in AWS EKS(Amazon Elastic Kubernetes Services):

• In a CircleCI pipeline, we lint the project's code, build a Docker image and deploy it to a public Docker Registry: Docker Hub
• Then in an AWS EKS cluster, we run the application
• Later, we promote to production a new app version using a rolling update strategy

All the project's tasks are included in a Makefile, which uses several shell scripts stored in the bin directory.

Project Tasks

Using a CI/CD approach, we build a Docker image and then run it in a Kubernetes cluster.

The project includes the following main tasks:

• Initialize the Python virtual environment: make setup
• Install all necessary dependencies: make install
• Test the project's code using linting: make lint
  • Lints shell scripts, Dockerfile and python code
• Create a Dockerfile to "containerize" the hello application: Dockerfile
• Deploy to a public Docker Registry: Docker Hub the containerized application
• Deploy a Kubernetes cluster: make eks-create-cluster
• Deploy the application: make k8s-deployment
• Update the app in the cluster, using a rolling-update strategy: make rolling-update
• Delete the cluster: make eks-delete-cluster

The CirclCI pipeline(config.yml) will execute the following steps automatically:

• make setup
• make install
• make lint
• Build and publish the container image

To verify that the app is working, write your deployment's IP into your browser using port 80, like http://localhost:80 or http://LOAD_BALANCER_IP:80 (according to your environment).

Alternatively, you can use curl: curl localhost:80 or curl LOAD_BALANCER_IP:80

CI/CD Tools and Cloud Services

• Circle CI - Cloud-based CI/CD service
• Amazon AWS - Cloud services
• AWS EKS - Amazon Elastic Kubernetes Services
• AWS eksctl - The official CLI for Amazon EKS
• AWS CLI - Command-line tool for AWS
• CloudFormation - Infrastructure as Code
• kubectl - a command-line tool to control Kubernetes clusters
• Docker Hub - Container images repository service

CicleCI Variables

The project uses circleci/docker orb, so to be able to build and publish your images, you need to set up the following environment variables in your CircleCI project with your DockerHub account's values:

• DOCKER_LOGIN
• DOCKER_PASSWORD

Main Files

• Makefile: the main file to execute all the project steps, i.e., the project's command center!
• config.yml: to test and integrate the app under CircleCI
• hello.app: the sample python/flask app
• Dockerfile: the Docker image's specification file
• hello_cluster.yml: EKS cluster definition file

The following shell scripts are invoked from the Makefile

• eks_create_cluster.sh: creates the EKS cluster
• install_eksctl.sh: installs the eksctl tool
• install_hadolint.sh: installs the hadolint linter(for Dockerfiles) tool
• install_kubectl.sh: installs the kubectl tool to control K8S clusters
• install_shellcheck.sh: installs the shellcheck(for shell scripts) linter tool
• k8s_cleanup_resources.sh: deletes services and deployments in a K8S cluster
• k8s_deployment.sh: deploys and exposes a service in the K8S cluster