Fedora and Red Hat and the vast majority of their derivatives use systemd.
OpenSUSE and SUSE and the vast majority of their derivatives use systemd.
In most production environments bar custom distributions used by Google & co. (and those are a minority), systemd is used, presumably with the systemd resolver.
I'm not sure I'd call all users of commercial Linux distributions "laymen" :)
Debian doesn't use nss-resolve(8) by default though. I'd be very surprised if RHEL or any other other distros suitable for a "production environment" were to take that step given systemd-resolved's current level of maturity.
(This is not a knock at systemd-resolved; I use it myself on most of my machines, but unless you want to spend a lot of time debugging weird resolver issues I would not recommend doing so until it has matured a bit more.
My only point was if you want "to take control" of your system you first have to look /etc/nsswitch.conf. If "resolve" is before "dns" your /etc/resolv.conf is mostly unused. Any article discussing /etc/resolv.conf without mentioning /etc/nsswitch.conf is very misleading for many users who happen to have systemd-resolved (probably because the distro decided to install and configure it)
Whether nscd is still installed by any distro by default I don't know. I haven't seen it running for years.
It would be fine if overall reliability increased, but it does not. I could write more reliable shell scripts for making WiFi or PPP connections than NetworkManager, which sometimes required a reboot before graciously allowing to connect again. Never happens on the command line.
Thank you for describing the status quo, I'll probably look into Slackware or OpenBSD now.