Ubuntu is the delusional cult of convenience that doesn't scale well to extremely large operations. Their whole premise was they released a completely free dist

Ubuntu is the delusional cult of convenience that doesn't scale well to extremely large operations. Their whole premise was they released a completely free distro that was identical to the commercial support version, as opposed to RHEL. If this is this case, then they've footgun fireworks factory fired themselves. Debian is fine, but RHEL-derivates are most reliable because they run at scale and have the most stable kernel going. CentOS (5 year lifecycle) is patched more often than Alma and Rocky (10 year lifecycle), so there's no "perfect" alternative free to RHEL. Large enterprises should be purchasing RHEL because most multibillion dollar companies run Cent or one copy of RHEL. Also, I wouldn't use RPMs because they muddy the boundaries between OS and app: nix, microdnf, or habitat where all of "your" stuff is isolated and vendored separately. At the boundaries, it's important to sanitize environment variables, PATH, and shebangs that could cause bleed through. When in doubt, create a minimal chroot environment (even within containerization, and use SELinux) because apps shouldn't be able to run wild over a system. Then s6 or a daemon tools-derivative non-init process supervisor that doesn't replace init is also important since Systemd is unreliable for real apps.