Homelab
Everything runs at home.
Hardware
Custom-built server on Arch Linux. Ubuntu VMs for validator workloads.
What is running
| Service | Purpose |
|---|---|
| Pi-hole/Technitium | Network-wide DNS ad blocking |
| Tailscale/Aperture | Mesh VPN linking all devices |
| Podman | Rootless container orchestration |
| Ansible | Config management for dotfiles and servers |
| Validator node | Quantum-resistant L1 protocol validator |
| Monitoring | Grafana + Prometheus + Loki |
| n8n | Self-hosted workflow automation |
| PostgreSQL | Primary database for projects |
Philosophy
Self-host what I can. If a SaaS exists that I can run myself, I run it.
Ansible everything. No manual server config. Playbooks or it did not happen.
Rootless containers. Podman, not Docker. Security by default.
Tailscale mesh. Every device reachable, nothing exposed to the internet.
Things I learned
Validator uptime is a different kind of reliability engineering. Financial penalties for downtime change how you think about monitoring.
Monitoring without alerting is wallpaper. You will not check the dashboard until something breaks. The alert has to find you.
Ansible idempotency is a superpower. Run the playbook, walk away, nothing changes that should not.
DNS is always the problem. Pi-hole is always the answer.
