Distributed Prime Counter

This is a fun project that I made to learn some stuff that I always wanted to get to know better. The application does not have a specific purpose other than that. I especially wanted to see how the scaling and load balancing behaves with K8s, which is why the ui displays the pods where the request went through.

Specifically, I used the following technologies:

• Kubernetes (using minikube for local development): Container Orchestration / Deployment
• Go: Programming Language
• gRPC: Service Communication
• Protocol Buffers: Communication Format
• Earthly: CI / CD
• React: User Interface
• Prometheus: Metrics Collection
• Grafana: Metrics Visualization

Architectual Overview

This is a very simplified overview of the architecture.

Getting Started

Using minikube:

# Setup minikube docker environment
minikube start
eval $(minikube docker-env)

# Build all services
earthly +all-docker

# Allows the access to NodePort and Ingress (Execute this in a seperate terminal)
minikube tunnel

# Deploy the application
kubectl apply -f deploy/

Now you should be able to access http://localhost:80 (for UI and API), http://localhost:9090/ (for Prometheus) and http://localhost:4000 (for Grafana).

Additional Notes

⚠️ Currently the Prometheus Deployment is running with a Service Account that has the cluster role cluster-admin assigned, which is not recommended for production!