Include some garbage collection information

[aluhrs13]

Jotting down a suggestion from a conversation, not sure where this might fall with the TAG gc principle.

If gc is part of a LoAF, ideally we could include how much of it was gc, but at least just a boolean to say "This LoAF was impacted by gc" would be helpful. Devs could use this to ensure they're looking at LoAF data more in their control, or understand if they see a pattern where they have a specific script/function hitting LoAFs that's impacted by gc more often than others and an indicator that they should look at better management of memory in that area.

Edit: After reading a bit more, does pauseDuration include gc?

[mmocny]

One alternative: We expose timing details for script that are >5ms, but what about exposing a totalScriptDuration which sums the durations for also scripts which are faster, and make this duration NOT include durations for work such as GC (not sure if all GC is created equal and worth ignoring, i.e. idle time GC vs GC during hot loops).

Then you could compare totalScriptDuration for the LoAF against the duration of the LoAF in order to get a general sense: how much of the time for this LoAF is accountable?

There might be many reasons beyond GC that affect LoAF durations, and where optimizing scripts wouldn't help.