/lab
Home lab
A personal lab for testing infrastructure and automation patterns before applying them to production analytics platforms. Not a showcase of tools. A place to break assumptions under real failure conditions.
[ what runs ]
The lab runs on Proxmox with Linux VMs and LXC containers. Networking is handled by OPNsense and OpenWrt with VLAN-based segmentation. All services run in Docker behind Caddy as a reverse proxy, with remote access through WireGuard and Tailscale. No internal services are directly exposed to the internet.
Automation runs through GitHub Actions, Ansible, and n8n. Observability is handled by Grafana and InfluxDB. Authentication is centralized through Authelia with SSO and MFA.
Home lab / Personal
[ principles ]
The lab is governed by constraints that keep systems predictable, operable, and recoverable under failure. These same principles carry over to production analytics platforms.
Environments are disposable
Provisioned, validated, and removed automatically. No long-lived test environments without clear ownership. Cleanup and drift detection are first-class concerns.
Recovery is tested, not assumed
Backups are incomplete until a restore has been performed. PostgreSQL backups are regularly restored into clean environments. Recovery time and failure modes are observed directly.
Access is segmented and minimal
Remote access via VPN only. Central authentication with SSO and MFA. Network segmentation using VLANs across admin, workload, and IoT boundaries.
Monitoring drives action
Metrics focus on latency, failures, and degraded behavior. Dashboards are built around what users experience. Alerts are tied to action, not noise.
What the lab deliberately avoids: manual recovery procedures, permanent environments without ownership, monitoring that cannot trigger an action, and automation that is clever but not predictable.
[ further reading ]
Lessons from a home lab
Read →Why infrastructure habits matter in analytics platforms.