1198211 - [e10s] Add measure(s) for swap-rate

Status: RESOLVED WONTFIX

(Toolkit :: Telemetry, defect)

Description

[Jim Mathies [:jimm]]

+++ This bug was initially created as a clone of Bug #1198209 +++

Part of the e10s release criteria work.

Comment 1

[Chris H-C :chutten]

:jimm, is this an existing measure that just needs a telemetry histogram (like the memoryResident stuff for bug 1198209?), or is this something that needs to be written from 0?

Comment 2

[Jim Mathies [:jimm]]

That's a question for Vladan, this requirement came out of his e10s release criteria research.

Comment 3

[Vladan Djeric (:vladan)]

(In reply to Chris H-C :chutten from comment #1)
> :jimm, is this an existing measure that just needs a telemetry histogram
> (like the memoryResident stuff for bug 1198209?), or is this something that
> needs to be written from 0?

This is something that will need to be written from scratch, and differently for different platforms

Comment 4

[Chris Peterson [:cpeterson]]

chutten, do we plan to implement telemetry measures for swap rate? Or should we WONTFIX this bug?

Comment 5

[Chris H-C :chutten]

If this is RAM<->disk swap rate, then it is a useful thing to have a measure of. Since this bug didn't have a clear explanation of what "measure(s) for swap-rate" was, and I don't know who to ask about it now that :vladan's left, I've found it easier to work on other, higher-priority work.

I'm sorry, I should've unassigned myself. I'll go ahead and do that now.

Comment 6

[Chris H-C :chutten]

From email from :njn: 

"- I'm skeptical that measuring the swap rate will lead to anything useful (though I'd be happy to be proven wrong :)  We already have a PAGE_FAULTS_HARD measurement that was added a long time ago and I don't think it's ever been useful."

That seems pretty clear to me that, not only is it not likely to be useful, we already have some coverage in the area supplying evidence that it is not likely to be useful.