Skip to main content
Architecture pages cover the parts of the stack that don’t belong to a single repo. They sit a level higher than the per-category sections and explain how data, code, and decisions move.

In this section

Data pipelines

The log and NetFlow ingest paths from the homelab edge to Splunk indexers. UniFi → HAProxy → Cribl Edge → Splunk.

AI development pipeline

The human → issue → AI agent → CI → review → ship flow. How Claude, Antigravity, and Copilot fit together.

Principles

  • One diagram per concern. Combining a deploy pipeline with a data pipeline shows nothing. Each diagram can stay small.
  • Diagrams over prose where structure matters. If a relationship is clearer as a graph, draw it.
  • READMEs own install and usage. The docs site owns architecture, philosophy, and cross-repo context.

Where to go next

Infrastructure

OpenTofu modules and the Proxmox stack.

Configuration

Ansible playbooks for host and app config.

Nix Ecosystem

Reproducible dev environments.

Observability

Logs and supported metrics go through Cribl to Splunk; trace routes to Langfuse and Phoenix are merged, not yet deployed.