Skip to main content

Homelab Overview

This homelab is designed as a production-ready Kubernetes cluster built on Raspberry Pi hardware, providing a cost-effective platform for learning and experimentation with enterprise technologies.

Architecture Overview​

graph TB
    Internet --> UniFi[UniFi Gateway]
    UniFi --> Switch[UniFi USW-Pro-24-PoE]
    Switch --> Pi1[Pi5 Master Node]
    Switch --> Pi2[Pi5 Worker Node 1]
    Switch --> Pi3[Pi5 Worker Node 2]
    Switch --> Pi4[Pi5 Worker Node 3]
    Switch --> Pi5[Pi5 Worker Node 4]
    Switch --> NAS[Synology NAS]

    Pi1 --> K8s[Kubernetes Cluster]
    Pi2 --> K8s
    Pi3 --> K8s
    Pi4 --> K8s
    Pi5 --> K8s

    NAS --> iSCSI[iSCSI Storage]
    iSCSI --> K8s

    K8s --> ArgoCD[GitOps with ArgoCD]
    K8s --> Prometheus[Monitoring Stack]
    K8s --> Apps[Applications]

Key Components​

Hardware Layer​

  • 5x Raspberry Pi 5: 16GB RAM each with active cooling
  • PoE+ Switch: Powers all Pi devices through single cable
  • NVMe SSDs: 256GB storage per Pi for fast I/O
  • Synology NAS: Network-attached storage for persistent data

Network Layer​

  • UniFi Gateway: Enterprise-grade routing and firewall
  • Network Monitoring: Real-time metrics collection
  • VLAN Segmentation: Isolated network segments for security

Kubernetes Layer​

  • Control Plane: Single master node with HA storage
  • Worker Nodes: 4x worker nodes for application workloads
  • CNI: Calico v3.31.3 via Tigera Operator for network policy and connectivity
  • Storage: Synology CSI for persistent volumes

Application Layer​

  • GitOps: ArgoCD for declarative application management
  • Monitoring: Prometheus, Grafana, and AlertManager
  • DNS: External-DNS with Cloudflare and UniFi providers
  • Ingress: NGINX Ingress Controller with TLS

Design Principles​

High Availability​

  • Multi-node cluster design
  • Persistent storage for critical data
  • Network redundancy where possible
  • Automated failover for applications

Observability​

  • Comprehensive metrics collection
  • Centralized logging with Loki + Promtail
  • Network performance monitoring
  • Infrastructure health monitoring

Security​

  • Network segmentation with VLANs
  • Certificate management with cert-manager
  • Secrets management with Sealed Secrets (GitOps-compatible encryption)
  • Regular security updates

Automation​

  • GitOps workflow for all deployments
  • Automated certificate renewal
  • Self-healing applications
  • Infrastructure as code

Benefits of This Setup​

  1. Learning Platform: Hands-on experience with enterprise technologies
  2. Cost Effective: Professional setup without enterprise hardware costs
  3. Scalable: Easy to add nodes or upgrade components
  4. Production-like: Real-world technologies and practices
  5. Energy Efficient: Low power consumption compared to traditional servers

Use Cases​

  • Development Environment: Test applications in Kubernetes
  • Home Automation: Self-hosted smart home applications
  • Network Services: DNS, monitoring, and network tools
  • Learning Lab: Experiment with new technologies safely
  • Backup Target: Reliable storage for important data

Performance Characteristics​

  • Total RAM: 80GB across cluster
  • Storage: 1.28TB NVMe + NAS expansion
  • Network: Gigabit Ethernet with PoE+
  • Power: ~100W total consumption
  • Availability: 99%+ uptime with proper maintenance

This architecture provides an excellent foundation for learning Kubernetes, cloud-native technologies, and modern infrastructure practices while maintaining reasonable costs and power consumption.