Monitoring
Grafana
Where I look when something feels slow: node health, disk pressure, backups, cert expiry.
Amsterdam · homelab
My photos, files, documents and home automation, all running on my own hardware. Everything below is what that actually looks like.
Services
Seven services doing the work I would otherwise rent. Each one replaced a subscription.
Software
Two Talos nodes, one cluster. Argo CD is the only thing that touches it, so merging a pull request is the deploy.
Hardware
No rented server anywhere in this. Ten units of aluminium extrusion, with 3D-printed trays for whatever never came with rack ears.
No power strip fits a 10-inch rack, so the back of it is printed too. It is a stack rather than a strip: every box brings its own module up with it, and nothing else in the rack changes.
Cut list, parts and the STEP files are open source — build your own
Open source
All of it is open source and on GitHub — one repository builds the cluster, one builds the rack.
The software
Every manifest Argo CD applies, the Talos node config, and the Renovate rules.
The hardware
The cut list, the parts and what they cost, plus STEP and 3MF for every printed mount.
Leave a star if you like what you see