If you are doing continuous profiling, you are probably using a low overhead stack sampling profiler rather than recording every method entry and exit.
That's a fair point. It really depends. For example, if you're recording method run times via an observability SDK at full fidelity, this could be an issue.