From (a LOT) of searching this seems setting a different active Activity on each monitor is longer be possible. But I'd like to get confirmation since it seems a rather big regression.

I have 2 physical monitors. They operate as one really wide virtual display. There does not seem to be any way to turn that off. It seems that would be necessary to do what I want.

I'd like to have a different Activity on each monitor. I tried messing around with Virtual Desktops which didn't sound like the right idea. As expected that just makes things more complex but doesn't get me any closer to having a different activity on each display. I read about KDE removing the ability to use different backgrounds on different monitors, although the reasoning didn't make any sense to me. It seemed to boil down to "we couldn't really figure out how to make it work consistently". I suspect my Activity issue was collateral damage of disabling this per-display setting.

So the question is fairly simple. Is there really no way to have different Activities on different monitors? Is this ever planned to be reinstated?

If it makes you feel better after not getting a response for over a year, I have never read poorer documentation for any system, or observed a software project with as much ADHD as KDE. Apparently any minor revision involves changing the entire underlying philosophy of the environment. Moreover--as far as I can tell--in cult-like fashion, once a new philosophy is adopted, and even if it will be overwritten within a matter of months, developers begin acting as if anything in violation of the newly adopted philosophy is blasphemy.

I don't remember different Activities on different monitors ever being implemented (except maybe via exotic x server hacks or some such . I do know they removed the feature for having different wallpaper+widgets on different Virtual Desktops on the same Activity, due to complexity problems similar to what you've described.

I do know that per-monitor Activities has been requested occasionally before, but just never got enough push to be implemented (though I'd guess that it also carries with it a significant amount of complexity as well)