Post #1129117

I just had a situation where, without #Codex, I would have been completely stuck. I have a single server hosting several projects. It’s good? Yes. It’s cheap. :-) But it has no backup, so I end up testing in production. I locked it down pretty tightly—for example, SSH access is only possible through my tailnet. Then I got the idea to configure an additional VPN, which turned out to be a bad move. Tailscale disconnected, and I was left without any access to the machine. The only rescue option was the noVNC console, which is completely inaccessible to me. Codex managed to get in through that inaccessible web console, rebooted the server, fixed the settings in rescue mode, rebooted again, and everything is working perfectly now. The whole thing took about an hour.