Crypto value web app

This repo includes the application code, Dockerfile, Helm chart and GitHub Actions workflow used to deploy a crypto currency value web app.

The application

The application is written as a python flask app, the code for which can be found in ./app.py, with the HTML file index.html in /templates.

If I had more time:

• I designed the code with variables for the coin_id and currency, with the idea that we could pass these as container environment variables in the Helm chart. Meaning that the crypto currency and the currency values could both be changed easily.
• I retrieved the crypto currency value every minute by using a refresh in the HTML. This feels like a bit of a hack to me, I know that this can be done in Javascript with jquery ajax, so with more time I would have tried to implement that.
• I would have added tests for the python code.

The Helm chart

The Helm chart lives in the ./chart directory. It deploys the application as service-a.

• I chose to use a LoadBalancer service to expose the service. I set this to HTTP as I tested it in minikube. This should use HTTPS, be secured by an SSL certificate and use a friendly DNS name, all of which could be done with more time/seemed out of scope for this exercise.
• In an enterprise environment we might want to instead use an ingress controller for this.

If I had more time:

• I generated the Helm chart using helm create, it's very basic with just enough changes to deploy the service. With more time I would refine it and add options such as the coin_id and currency container env vars as mentioned above.

The GitHub Actions workflow

The Github Actions workflow is defined in .github/workflows/build-and-deploy.yaml.

Workflow steps:

• build docker image
• scan the image for vulnerabilities using the action snyk/actions/docker@master
• lint the helm chart
• test installing the helm chart
• there is then an approval gate created using the environment argument. This allows us to check the results of the vulnerability test and helm linting and testing before proceeding.
• once approved the docker image is pushed to ECR, and the helm chart is deployed

If I had more time:

• For the purposes of the exercise I wanted to show all of this in one workflow. It might make sense to separate these jobs into different workflows and to deploy the Helm chart with a GitOps CD tool such as ArgoCD which will maintain the kubernetes objects state to match what is in the git repo.