Currently, heap profiler stacks don't have any meaningful identification to tie them back to the thread/conn responsible. Doing so would help tie back to the specific operations that cause anomalous memory usage.