That’s basically what I’m thinking about my homelab. Almost nothing has a defined fail over, most things run as a single instance on the very same (used, old) hardware but my uptime is way higher than some business critical apps at work LOL