Analysing allocator CPU & memory issues

Nobody has claimed this yet.

I have a problem with `mimalloc` and I'm a bit unsure how to approach it. My application consumes a lot of memory using `glibc` allocator, so I switched to mimalloc. It works wonders, except when it doesn't. Sporadically I can see following pattern in my application:

I know for a fact that there is **_no user code_** in the application that can spike into 100% CPU and 5x memory usage. I am pretty sure it's an allocator (and this graph looks very different on glibc allocator). This only happens in production environment on real data, and I haven't been able to create a MRE to showcase the isssue.

Is there any way I can tell why/how this is happening, if I am unable to shrink the issue to the MRE? Maybe I can create a worker to log some of the internal mimalloc structs, so when this happens we have some numbers to work from? At this point I'm a bit lost because I haven't had such problems before and I am not sure how to approach it.

Start with the production CPU and memory graph and compare the same workload under mimalloc and glibc, since no minimal reproducible example or source location is provided. Gather allocator-internal measurements during an occurrence and document the application conditions; done means identifying evidence for the spike or narrowing it to a reproducible allocator issue.

Written by the indexing model from the issue text.