I Built a Production-Ready Kubernetes Cluster for $0 (Real Bug Included)

Welcome to practical AI-powered DevOps. On this channel, I help engineers and technology professionals automate their work using AI—building faster deployments, smarter infrastructure, reliable CI/CD pipelines, and workflows that can monitor and repair themselves. You'll find hands-on tutorials, real-world experiments, comparisons, and career guidance covering AI agents, DevOps automation, AWS, Kubernetes, Docker, Terraform, GitOps, cloud computing, platform engineering, and more. — Managed Kubernetes costs around $70/month before you deploy a single pod. In this video I build the exact same production patterns — ingress, TLS, autoscaling, monitoring, and zero-trust networking — entirely on my own laptop, for $0. I didn't edit out the part where it broke. Early on, my ingress controller hits a real Kubernetes scheduling bug that trips up almost everyone the first time they set up a multi-node cluster with kind — and I diagnose and fix it live, on camera, the same way you'd troubleshoot it in production. What's covered: ✅ Multi-node cluster setup (kind) ✅ Ingress controller — including a real bug, fully diagnosed and fixed ✅ TLS termination with cert-manager ✅ Horizontal Pod Autoscaling (HPA) with real metrics ✅ Full observability with Prometheus + Grafana ✅ Zero-trust network policies, proven with a live denial test Total cost: $0.00 Chapters: Intro — why managed Kubernetes costs money and this doesn't Cluster setup Ingress layer + the real bug TLS / cert-manager Autoscaling (HPA) Monitoring (Prometheus + Grafana) Network policies Recap + what's next 📦 Manifests + bug write-up: https://github.com/Mamadoubalde/zero-... This kicks off a series where I hand real DevOps workflows to AI agents and measure exactly what happens — time saved, what breaks, what holds up. Subscribe so you don't miss what's next. #Kubernetes #DevOps #CloudComputing #k8s #SRE #AIagents