Retro Video Game Exchange

A Kubernetes-orchestrated microservices project from my distributed systems course, load-tested with k6 and autoscaled via HPA.

KubernetesDockerKafkaNginxPrometheusk6JavaScript

Product Preview

Project Screenshot

The Problem

A multi-service exchange needs to stay responsive under sudden traffic spikes without manually resizing pods or losing message flow between services.

The Solution

Deployed the Retro Video Game Exchange on Kubernetes with Kafka messaging, Nginx, and Prometheus, then ran a k6 load test (100 VUs) while watching the Horizontal Pod Autoscaler scale the API from 1 to 10 replicas under CPU pressure and back down after the spike.

Technical Challenges

  • Configuring Kubernetes deployments, resources, and HPA targets so autoscaling could react to CPU load
  • Wiring Kafka bootstrap servers and supporting services (Kafdrop, Prometheus exporters) into the cluster
  • Writing and running k6 load tests that drove ~39k iterations against the live API
  • Observing and validating scale-up/scale-down behavior under watch with kubectl