fzo-kube

Install K3D

To run the services on your own machine, you can install K3D.

To install it, run curl -s https://raw.githubusercontent.com/k3d-io/k3d/main/install.sh | bash.

For this assignment, a node port is needed, so you can run this command:

$ k3d cluster create mycluster -p "8082:30080@agent:0" --agents 2

This will create a cluster exposing a potential NodePort on port 30080, mapped to the host's port 8082.

Application

The simple application made for this assignment is a last timestamps retriever.

When someone loads the frontend page, it registers when the page was accessed and stores it in the database. The database then provides the list of the last 5 timestamps to have been written in.

Technologies used:

• frontend: Python with Flask
• api: Deno
• database: PostgreSQL

Deployment

The images are built using Github Actions and stored in the Github Container Registry (ghcr).

Description

WARNING: this deployment is not made for production and is not secure

As per requested in the assignment, the database and the API have been placed in two different namespaces. As such, there are two secrets with contain the same values, both in the api/ and the database/ folders of the kubernetes deployment. For a production environment, secrets wouldn't be committed and would be shared in order to ease key rotation and other features.

The frontend is accessible by a NodePort and contacts the API though its ClusterIP service. The API then contacts the database through its ClusterIP service.

All three elements feature ConfigMaps (and Secrets if needed) to ease configuration and separate resources.

The database also possesses a PersistentVolume and a PersistentVolumeClaim in order to ensure data retention.

Instructions

First, load up the namespaces using this command:

$ kubectl apply -f kubernetes/namespaces/

Then, apply all elements:

$ kubectl apply -f kubernetes/database/
$ kubectl apply -f kubernetes/api/
$ kubectl apply -f kubernetes/frontend/