[reported from users of Leyden EA]

I used https://github.com/hatoo/oha to simulate a production workload with `oha -z 100s http://localhost:8080/owners`, generating a lot of requests to this single endpoints.

It took a very long time to generate the archive, like 100 second on my powerful MacBook Pro after applying some workload. That creates issues like containers being killed before the archive is created and requires kill timeout customization that will not be possible on all platforms. Is it a known limitation with some potential upcoming optimization possible?

I expect typical production workload to exercise much more endpoints with even more requests, so I suspect there may be a need for optimizations to handle real life use cases. Maybe you can limit how much profiling data are collected/analyzed?