this will allow annotations to collect in the new bugs and reduce confusion with the old bugs.

crashes are pretty much done, we need a plan for leaks.

I see a few types of leaks to figure out:
Intermittent LeakSanitizer | leak at NewNonOwningRunnableMethod, mozilla::MediaTransportHandlerSTS::Destroy_s, applyImpl, apply
Intermittent ShutdownLeaks | process() called before end of test suite [only use for classification if there is no test failure directly mentioned after it]
Intermittent leakcheck | tab 292361 bytes leaked (APZEventState, ActiveElementManager, AsyncFreeSnowWhite, AttributeStyles, BackstagePass, ...)

I will ignore AddressSanitizer and ThreadSanitizer bugs as part of this work.

The big fish here are leakcheck and LeakSanitizer

I guess what is useful in the summary and the bug information. Do we care about what test (or manifest) was run to generate this?

Are there ways we can bucket this to be more beneficial?

:decoder, I am starting with you, possibly you have some ideas, feel free to redirect to someone else.