Project

Deploying a Flask application container to AWS EKS and viweing the webpage via the Elastic Load Balancer. In this project I am working with:

• AWS
• CircleCI to implement Continuous Integration and Continuous Deployment
• Linting using Makefile
• Docker
• Kubernetes clusters
• Flask Accessing the deployed app on elb.

Steps in Completing the Project

Prerequisites

• AWS account
• A Circleci Account. In this configure:
  • Add the AWS credentials as environment variables. Configure AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY and AWS_DEFAULT_REGION
• Your App to deploy

Tools

1. Circleci
2. Amazon Elastic Kubernetes Service (EKS)
3. Dockerhub (Alternatively, you can push your image to AWS ECR)
4. Kubernetes
5. CloudFormation
6. eksctl

Steps

1. Fork the project to your Github Account
2. Add the AWS credentials as environment variables. Configure

• AWS_ACCESS_KEY_ID,
• AWS_SECRET_ACCESS_KEY
• AWS_DEFAULT_REGION

3. Complete your Dockerfile to build the image of your app

4. Add Makefile for linting

5. Complete your .circleci/config.yml. Implement Build and deploy jobs with CircleCI kubernetes and AWS EKS orbs.

   orbs:
         aws-eks: circleci/aws-eks@2.1.1
         kubernetes: circleci/kubernetes@0.4.0
         

   • Using the Makefile you can add more linting stages.
   • Build and push your docker image to Dockerhub. Example:

   build-and-push-image:
   machine: true
   steps:
        - checkout
        - run: |
           docker build -t greentropikal/covid19-analysis:latest .
           echo $DOCKERHUB_PASSWORD | docker login -u $DOCKER_USER --password-stdin
           docker push greentropikal/covid19-analysis:latest

   • In the workflow create the cluster and all the required VPC-related resources to run Kubernetes on AWS.

      
       create-cluster:
           docker:
           - image: << pipeline.parameters.executor>>
           steps:
           - aws-eks/install-aws-iam-authenticator:
               release-tag: ''
           - aws-eks/create-cluster:
               cluster-name: << pipeline.parameters.cluster-name >>
               node-type: t3.medium

   • Some of the resources created using this simple command are NATGateway, eks cluster, ec2, loadbalancer and so many more. Check out the template here for more.

   • Use the deploy-application command to deploy your application to eks. This is where you declare your deployment strategy. The main commands are:

       
           - aws-eks/update-kubeconfig-with-authenticator:
               cluster-name: << pipeline.parameters.cluster-name >>
               install-kubectl: true
               aws-region: << pipeline.parameters.aws-region >>
           - kubernetes/create-or-update-resource:
               get-rollout-status: true
               resource-file-path: deployment/deploy.yml
               resource-name: deployment/covid19-analysis
           - kubernetes/create-or-update-resource:
               resource-file-path: deployment/app-service.yml
     

   • Test your application to check if everything is working out fine.

6. Push your code to github and the pipeline will do the rest.

Links

• circleci/aws-eks orb.