Tracing memory leaks with `tracemalloc.take_snapshot()` is extremely slow

Question

I have a Flask service that leaks memory. I use the built-in tracemalloc library to trace it, similarly to this guide. However, my service, when it loads, takes about 9G of