I am showing my playtesting summary here, but this but no means all the QA for the project. QA started way earlier - when planning the classes I would program.

In my experience almost no amount of testing can fix a badly planned piece of software with a lot of "mysterious" dependencies. In my project weren't such issues, so bugfixing was a fast and straight-forward affair. 

Another problem is of course performance on webGL - since according to the profiler the program logic wastaking negligble time compared to the rendering I switched to URP and tuned down the quality settings, but still it is a bit weak on the framerate side.