📄️ Study Case Overview
Act as: Senior DevOps / Platform Engineer mentoring a junior engineer.
📄️ Container Runtime (containerd)
Production Purpose: Every Kubernetes node needs a container runtime. containerd is the CNCF-standard runtime used by all major managed Kubernetes services (EKS, GKE, AKS). Choosing it bare-metal mirrors exactly what cloud providers run under the hood.
📄️ Bootstrap Kubernetes Cluster with kubeadm
Production Purpose: kubeadm is the official Kubernetes cluster bootstrapping tool. It handles control-plane initialization, TLS certificate generation, etcd setup, and node join tokens — the same steps cloud providers automate for managed Kubernetes (EKS, GKE, AKS).
📄️ Calico CNI Networking
Production Purpose: Kubernetes itself does not implement pod networking. It delegates to a CNI (Container Network Interface) plugin. Calico is the most widely deployed CNI in production — used by major banks, telcos, and cloud-native platforms — because it provides scalable, policy-aware networking using BGP routing.
📄️ MetalLB Load Balancer
Production Purpose: In cloud environments (AWS, GCP), a LoadBalancer Service automatically provisions an external load balancer. On bare-metal, this doesn't exist. MetalLB fills that gap — it gives bare-metal Kubernetes clusters a real external IP for Services, exactly how cloud providers do it.
📄️ Ingress NGINX Controller
Production Purpose: Running one LoadBalancer Service per application is wasteful and expensive — each Service consumes an external IP. Ingress solves this by using a single external IP (from MetalLB) and routing traffic to multiple backend services based on hostname and URL path. This is the standard pattern used by every production Kubernetes platform.
📄️ Persistent Storage (PV, PVC, StorageClass)
Production Purpose a misconfigured PV is the #1 cause of data loss incidents in Kubernetes.
📄️ Laravel Production Deployment
Production Purpose PHP-FPM + Nginx sidecar, MariaDB with persistent storage, Redis for cache/sessions, ConfigMaps for environment config, Secrets for credentials, and proper health probes.
📄️ Monitoring with Prometheus & Grafana
Production Purpose: "You cannot improve what you cannot measure." Prometheus is the CNCF standard for Kubernetes metrics — it powers monitoring at Google, Netflix, Cloudflare, and virtually every production cloud-native platform. Grafana visualizes those metrics with dashboards that surface problems before users notice them.
📄️ GitOps with ArgoCD
Production Purpose: GitOps means Git is the single source of truth for your cluster state. You never kubectl apply directly in production — instead, you push to Git, and ArgoCD automatically detects and syncs the changes. This gives you full audit trail, rollback history, and drift detection for free.
📄️ Security Hardening
Production Purpose: A default Kubernetes installation is NOT secure. Pods run as root, all pods can talk to all other pods, Secrets are only base64-encoded, and any user with kubectl can do anything. Security hardening closes these gaps systematically — following the principle of least privilege at every layer.
📄️ Failure Simulation & Troubleshooting
Production Purpose: The only way to build confidence in a production system is to break it deliberately and practice recovery. This phase simulates real production failures — node crashes, OOM kills, database corruption, network splits, and misconfigured rollouts — and teaches you the systematic approach that SREs use to diagnose and fix them.