1599908 - [meta] Determine sources of noise in visual metric results

Status: RESOLVED INACTIVE

(Core :: Performance: General, task, P3)

Description

[Andrew Creskey [:acreskey]]

We have identified (and fixed) numerous sources of variable page load results in Bug 1502138.

This bug will track sources of noise that specifically affect the visual metrics, such as SpeedIndex.

Comment 1

[Andrew Creskey [:acreskey]]

Attachment: SpeedIndex measured through a webpagereplay recording on geckoview example.

A recent example shows that visual metrics also return both outliers and a high degree of variance.

Comment 2

[BugBot [:suhaib / :marco/ :calixte]]

The meta keyword is there, the bug doesn't depend on other bugs and there is no activity for 12 months.
:kmoir, maybe it's time to close this bug?

Comment 3

[Kim Moir [:kmoir] ET]

I talked to Andrew and he would like to keep it open.

Comment 4

[BugBot [:suhaib / :marco/ :calixte]]

The meta keyword is there, the bug doesn't depend on other bugs and there is no activity for 12 months.
:plawless, maybe it's time to close this bug?

Comment 5

[Andrew Creskey [:acreskey]]

We haven't specifically looked into this issue and I believe there is the potential for us to find improvements in our rendering pipeline to improve both noise and visual metrics.
So I think there's value in keeping this bug open.

Comment 6

[BugBot [:suhaib / :marco/ :calixte]]

The meta keyword is there, the bug doesn't depend on other bugs and there is no activity for 12 months.
:fdoty, maybe it's time to close this bug?

Comment 7

[Frank Doty [:fdoty]]

:acreskey Is this a bug which still needs to be tracked, and should this be a meta bug?

Comment 8

[Andrew Creskey [:acreskey]]

Hmmm.
I think that history has spoken and this bug was not particularly useful in collecting bugs related to noise in visual metrics.
I'm happy to close it.