Content System
This page defines the repeatable system for growing the wiki without turning it into a folder of disconnected notes.
Page Types
| Type | Purpose | Best Format |
|---|---|---|
| Track page | A durable domain such as AI architecture, platform engineering, or hardware | Overview, topic list, flagship links |
| Project page | A complete system or build | Architecture layers, diagrams, decisions, runbooks |
| Build log | A dated implementation journal | Context, attempt, failure, fix, learning |
| Decision record | A durable architecture choice | Context, constraints, options, decision, consequences |
| Visual reference | A diagram-first explanation | Mermaid, Excalidraw export, SVG/PNG, notes |
Frontmatter Standard
Use consistent metadata on every substantial page. This improves search, navigation, future automation, and content audits.
title: Private Compute Cluster
description: OpenStack, Ceph, KVM/QEMU, and SDN architecture notes.
slug: /engineeringnexus/projects/private-compute-cluster
tags: [OpenStack, Ceph, SDN, KVM, Platform Engineering]
status: active
level: advanced
updated: 2026-05-24
Writing Standard
Every strong page should make the engineering visible:
- Define the system or idea in plain language.
- State the constraints and why they matter.
- Show the architecture with at least one visual representation.
- Explain tradeoffs, decisions, and failure modes.
- Capture what changed after implementation.
Visual Standard
Use the simplest visual that communicates the system accurately.
| Need | Use |
|---|---|
| Fast flow or dependency map | Mermaid |
| Rough system exploration | Excalidraw |
| Portfolio-grade reference | SVG or PNG |
| Tradeoff comparison | Table or decision matrix |
| Durable architecture choice | Decision record |
Search Tags
Prefer durable tags over temporary moods. Good tags include technologies, domains, patterns, and outcomes:
OpenStack,Ceph,KVM,QEMU,SDNAI Architecture,Agents,RAG,EvaluationPlatform Engineering,SRE,Kubernetes,CI/CDDrones,IoT,Embedded Linux,Edge ComputingAlgorithms,CAT,Quant,DILR,VARC