It is interesting how long server uptime used to be considered a good thing but now the shorter the average runtime the better as it indicates the system was designed to tolerate failures.
No, a long uptime is evidence of the crime of not sacrificing sufficiently on the altar of security. You must be constantly interrupting service for security patches, firmware updates, release upgrades, etc. so your users are constantly reminded that security is your paramount concern.
Patching vulnerabilities IS important and well designed systems can keep running while one server reboots.
Linux may get a kexec handover protocol some years from now. You can reboot the kernel and your userland daemon seperately :)