Steps for setup

Required packages

• minikube
  • brew install minikube (installs kubectl[kubernetes-cli], docker-runtime and docker-daemon as well)
  • brew install hyperkit
• jq (Command-line JSON processor)
  • brew install jq

Starting the application

bash setup.sh => outputs the ingress ip-address

Updating the application

On every commit, Github workflow updates the image in docker hub.
To update the running pods, run:

• bash deploy.sh

Application architecture/strategy

• The application is a basic flask-app that runs on a local minikube cluster
• It runs as a docker container and being a flask app listens on port 5000 by default
• It exposes 2 endpoints:
  • [GET] / => "Well, hello there!"
  • [GET] /healthcheck => "OK"
• It is dockerized and the image is available in docker hub:
  • docker pull amanmahajan26/flask-app
• The build process is automated using github workflow, i.e, every commit or PR against master branch triggers the build pipeline which updates the docker image and pushes it to docker hub. The username and password for docker hub are configured in the github secrets.
• The container runs inside a pod, the smallest computing unit in Kubernetes
• To maintain high availablity, we maintain 3 replicas of the pod which are managed by a Deployment
• The Ingress resource does the load balancing
• Service component is responsible for providing a single entrypoint for the pods
• The container also have liveness and readiness probes configured to ensure that containers are healthy and ready before serving traffic

Connecting to application

Launch the ingress ip-address received as output of running setup.sh script in a browser