App Insights Profiler doubles memory consumption

Hi,
I have a question regarding App Insights Profiler.

We've configured our app according to this tutorial and locally (run from VS or in a Docker container) everything works fine (there are API metrics in Portal, profiling sessions, code optimizations etc.), it consumes a little bit more memory, but I see no spikes. When deployed to the dev environment (kubernetes cluster) the memory usage goes bananas during profiling sessions and the app consumes twice the usual amount of memory. Also, it seems there are way too many sampling profiling sessions, it is not rare to see 20 per hour.

The documentation says that the sampling profiling sessions are scheduled once an hour and that when Profiler is active the overhead is up to 15%.

I'm not able to find any relevant info to this. Thanks for any additional questions or help in advance.