1050246 - 6.79% OSX 10.8 Paint regression reported on Aurora (v.33) but originally seen on m-i ~ June 24

Status: RESOLVED WONTFIX

(Testing :: Talos, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

This is difficult to figure out as we never received a graph server alert originally, therefore it was not brought to attention.  Why we didn't receive an alert, I have no idea, it is a clear looking regression.

Here is a graph showing the 90 days of history:
http://graphs.mozilla.org/graph.html#tests=[[82,52,24],[82,1,24],[82,63,24]]&sel=none&displayrange=90&datatype=running

Focus on a few data points:
Jun1 - values drop for both inbound and firefox branches
Jun9 - merge firefox -> aurora (no issues as we had a lower value)
Jun23 - values raise again for inbound and firefox branches
July21 - merged firefox -> aurora, we see the jump

looking over previous alerts and bugs, I have no indication of what went wrong.  What makes this harder is that the data is too old so that retriggers on inbound are not an option.

This is the only paint regression we have seen from the aurora uplift, and for osx, we have this paint regression and some smaller dromaeo regressions.

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

we have already shipped this.