home-ops

One Talos node, one Flux repository, no configuration done by hand.

Decommissioned

The cluster was decommissioned on 2026-09-29 and no longer runs. These pages are kept as a historical record. Selected services (Umami, Cairn, the Gatus checks, the DNS fallback and slskd) are being rebuilt on a VPS platform in the private fobiat/vps-ops repository, tracked there as issue #43.

If you are looking for something specific:

• Setting this up from nothing: Bootstrap
• Adding an app to the dashboard: Homepage
• Adding a health check for an app: Gatus
• What's actually backed up: Cluster backup
• Analytics: Umami
• Publishing something to the internet: Exposure
• The disk died: Restore
• Upgrading Talos or Kubernetes: Upgrade (or via tuppr)
• Something is broken: Troubleshooting
• Why is it like this: Decision records

The shape of it

Flux watches this repository and reconciles it into the cluster. Nothing else writes to the cluster. A change is a pull request, and the pull request carries a rendered diff of what will actually change.

Talos owns the node. Its configuration is a file in talos/, applied through an API. There is no SSH and no package manager, so there is no way for the node to drift from what is written down.

Access is Tailscale first. Services live on *.lab.fobiat.dev and resolve only inside the tailnet or on the local network. Making something public is a deliberate, documented act rather than a default.

Honest limitations

This is one machine. Upgrades are outages. There is no failover, no quorum, and no second copy of anything running. The single node decision record lists exactly what that costs and what is deliberately not attempted because of it.

Backups currently go to a second disk in the same machine, which protects against deleting a volume and against nothing else. See Cluster backup for what's actually covered. Offsite is planned and is a prerequisite for the hardware upgrade.