Skip to content

Latest commit

 

History

History

README.md

Architecture Decision Records (ADRs)

This directory documents the foundational architectural and design decisions governing the stefanutc1/infrastructure platform. Each record outlines the architectural context, decision taken, status, alternatives considered, and operational consequences.

Governance & Process

  1. Immutability: Once an ADR is accepted, it reflects an architectural consensus. If a decision changes, a new ADR must supersede it.
  2. Resource Awareness: Every architectural decision must account for the physical hardware reality of the homelab fleet (12GB RAM hypervisor, 2GB NAS, 4GB edge compute node).
  3. No Fake Enterprise / Anti-Cargo-Culting: Decisions prioritize lightweight, highly integrated solutions over bloated enterprise suites that degrade reliability.

ADR Index

ADR ID Title Status Date Primary Driver
ADR-0001 Proxmox VE as Primary Bare-Metal Hypervisor vs. Pure Bare-Metal Kubernetes Accepted 2026-09-20 Virtualization Density & Mixed OS Support
ADR-0002 LXC Unprivileged Container Density for Core Microservices Accepted 2026-09-20 RAM Optimization (<100MB per microservice)
ADR-0003 Virtualized OPNsense Dual-Perimeter Gateway with Bridge Transit Accepted 2026-09-21 Deep Packet Inspection (Suricata) & Zero Trust Routing
ADR-0004 Two-Tier Storage Architecture: Local NVMe local-lvm & Remote ZFS NAS Accepted 2026-09-21 Low-Latency I/O vs. Fault-Tolerant Bulk Capacity
ADR-0005 Ingress mTLS / OAuth2-Proxy & On-Demand Active Directory Research Lab Accepted 2026-09-22 Zero-Bloat Identity vs. Academic Research Depth
ADR-0006 GitOps Zero-Plaintext Secret Hygiene & SOPS/Age Ingestion Accepted 2026-09-22 Leak Prevention without Heavyweight Vault Cluster
ADR-0007 Single-Node Lightweight Kubernetes (k3s/k0s) on Legacy Hardware Accepted 2026-09-22 Edge Workload Isolation on AMD Athlon II Node
ADR-0008 ELO Multi-Tier Cascade Routing & L0–L3 Tool Execution Gatekeeper Accepted 2026-09-23 Local GPU Inference (GTX 1050 Ti) & Non-Destructive AI Governance