I have been a Tailscale user and fan for a few years now. I rely on my Tailnet extensively to connect to my homelab and other devices, securely from anywhere. Lately, I’ve run into a limitation: making services highly available (HA) over Tailscale isn’t straightforward.For example, I have a highly-available MinIO cluster across three nodes, and I also wanted a reliable way to proxy into my UniFi console from multiple nodes (all running Traefik). Ideally, both of these endpoints (and a few more) needed to stay available even if one machine failed.