Tooltips

Oh, it's super old and shouldn't be accounted for anything.

I think the biggest issue with local testing is that the client and server are both working on the same resources and performance hits tend to show up a lot worse in those cases. The only issue pure clients would really notice are small spikes and a gradual increase in client memory usage. Since SS13 reboots every round there's not a ton of time for the issue to compound too greatly, but if you took it to something that runs for days or weeks at a time you'd have clients who have been logged in for a long time starting to notice their clients running at more CPU than it ever should and using a lot more memory than it should.

Eternia actually had the same problem, our admin panel buttons were just that, GDI buttons on top of the map that were shown/hidden as needed, the problem was that those buttons existed for every client, and despite being hidden they accounted for a massive drop in performance because they always existed and were on top of the map (didn't help it was updating beneath it, but that really just makes the problem glaringly more obvious).

Once I moved those to their own little window (a bandaid until I can make a more proper replacement panel) client-side performance and memory usage jumped dramatically. Then I noticed I had a browser window sneaking around doing the same thing, removed that, and bam, another massive performance jump.

I think the most major issue is that the browser windows aren't actually being destroyed, so they can account for an increasing curve of performance loss the more they're used.

It also never frees up the memory the windows are using, if you watch it in task manager and use the hover tip you can watch the numbers steadily increase but never decrease (I don't think there's actually a way to delete a created window)