Infrastructure engineering, homelab projects, and lessons learned in the field
Recent lessons from running my homelab day to day: guardrails for AI-assisted operations, root cause analysis, and turning incidents into tooling. Each post ends with what the lesson means for customers.
How I run my homelab with Claude Code as a working partner, and the guardrails, 1Password secrets handling, and written rules that make it safe. Every rule came from something that broke.
Read MoreTwo incidents where the obvious suspect was wrong: a reboot that hung for hours, and a network outage that looked like a failing NIC. Root cause, the safe-reboot tool that came out of it, and lessons.
Read MoreA three-part series documenting a major infrastructure migration in my homelab, covering port conflicts, DNS management, and troubleshooting. This project involved migrating Traefik to a dedicated host, updating 31 DNS records, and implementing proper separation of concerns.
AI-Assisted Workflow: This series was built with CLI-based AI tools (Gemini CLI and CodeX at the time) with context maintained in markdown files. I’ve since standardized on Claude Code for all infrastructure and writing work. The approach still holds: markdown files give continuity across sessions and serve as living infrastructure documentation.
How I resolved a port 443 conflict between Traefik and AdGuard Home by migrating my reverse proxy to a dedicated Docker host. Covers the decision-making process, implementation challenges, and architectural improvements.
Read MoreManaging 31 DNS record updates, resolving CNAME conflicts, and purging 35 obsolete records. Includes working with the Cloudflare API, Docker cleanup operations, and implementing log rotation policies.
Read MoreDebugging a redirect loop that appeared after the migration. Covers the diagnostic process, implementing a multi-part fix with DNS CNAMEs, Traefik middleware, and handling self-signed certificates.
Read More