Closed Bug 545741 Opened 14 years ago Closed 14 years ago

Send Fennec perf regression emails to public location

Categories

(Release Engineering :: General, defect)

Product:
Release Engineering
Component:
General
Platform:
ARM
Maemo
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: mozilla, Unassigned)

References

Details

Detecting regressions came up as a top priority in the Fennec 1.0 post-mortem.

These have been going to me directly so I could help prevent spamming anyone with potentially confusing information (if they were too noisy to be useful, for example), but these seem like they're not too spammy so they should go somewhere public.

I'd say moz.dev.tree-management as well, so Firefox and Fennec developers can see these.

They currently look like this, though with links:

From: Nobody <nobody@cruncher.build.mozilla.org>
Date: Sat, 30 Jan 2010 14:17:05 -0000
To: aki@mozilla.com
Subject: Talos Regression: Tp4 increase 22.19% on Nokia n810 mobile-tracemonkey

Regression: Tp4 increase 22.19% on Nokia n810 mobile-tracemonkey

Previous results: 27156.3 from build 20100128022652 of revision 269af0364214 at 2010-01-28 11:40:10 on maemo-n810-53 run # 0

New results: 33182.3 from build 20100128022652 of revision 269af0364214 at 2010-01-28 21:40:14 on maemo-n810-41 run # 0

Suspected checkin range: from revision 269af0364214 to revision 269af0364214
cc'ing general@fennec.bugs.

Also, I marked bug 545472 as a dependency, as the revision linked to and specified is the mobile-browser revision, but the link goes to the tracemonkey repo.  I'm guessing that the mozilla-* version is going to be more granular and potentially more useful to track than the mobile-browser version (we can currently only track one in graphs; extending that is significantly beyond the scope of this bug).
Also, the email above appears to be the same buildid ?  Certainly possible there's device-specific noise here.
The Tp4 graph is questionably useful, as most runs crash mid-run and the resulting graph is very very noisy. http://3.ly/1Xwo

The Tdhtml graph, however, looks like a valid regression. http://3.ly/i2v1

I think I should fix the revision (mobile-browser -> mozilla-* revision, bug 545472) then ask catlee to send these to moz.dev.tree-management.
Ok, I think bug 545472 might take longer; in the meantime I've gotten two more regression emails (DHTML increase 11.24% on m-c, 12.03% on tm).

Catlee: could we send these emails to moz.dev.tree-management along with the Firefox desktop talos emails?  Assuming you have no objections.
Assignee: aki → catlee
Assignee: catlee → nobody
Done
Status: NEW → RESOLVED
Closed: 14 years ago
Resolution: --- → FIXED
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.