The second print queue disappearing implies that it is being auto-setup by cups-browsed. It should reappear when cups-browsed is restarted. This part of remote configuration is working correctly and as designed.
The first queue is not set up by cups-browsed but by the GTK printing sub-system. I am guessing that the screenshot shows a GTK application. (You did not say what the application is).
GTK gets remote printers all by itself. It does not use CUPS but deals directly with the printer. My guess is that when CUPS is stopped (
systemctl stop cups) the first entry still does not disappear.
The issue of GTK not using the correct CUPS APIs may be seen as a bug. There is nothing you can do about the first entry without severely crippling your whole printing system, which is what some of the advice in @User24601's link advocates. Users will have to live with it until the GTK maintainers decide to fit in with what CUPS provides.
Informative links are here and here.