What's the best way to profile Kedro pipeline? I want to reduce the end-to-end time.

we've been lucky with pyinstrument! see for example