By the way, I'm having fun looking at some uptime data. Here's some:
%up 99.996 | since Wed Nov 22 10:51:03 2023
%up 99.998 | since Fri Jul 29 09:35:57 2022
%up 99.999 | since Wed Apr 6 14:48:49 2022
%up 99.931 | since Thu Sep 7 08:48:55 2023
%up 99.989 | since Sun Mar 20 18:06:40 2022
%up 99.994 | since Thu Dec 9 17:10:22 2021
In other words, they were only offline for updates that needed a reboot.
These are all leased production servers located in Europe, running FreeBSD with jails and VMs.
No need for Kubernetes or the cloud to get great uptime!