New issue
Advanced search Search tips

Issue 811602 link

Starred by 1 user

Issue metadata

Status: WontFix
Owner: ----
Closed: Mar 2018
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

4.3%-8.9% regression in thread_times.key_mobile_sites_smooth at 535709:535814

Project Member Reported by sullivan@chromium.org, Feb 13 2018

Issue description

See the link to graphs below.
 
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Feb 13 2018

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=811602

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=0cf81e3e1f990feac573181085fd0675707b6463caef7870d2ac6990aa85b948


Bot(s) for this bug's original alert(s):

android-nexus5
android-nexus5X
android-webview-nexus6
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Feb 13 2018

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/12e72f6d840000

Comment 5 by sullivan@google.com, Mar 13 2018

Trying another bisect.
Project Member

Comment 6 by 42576172...@developer.gserviceaccount.com, Mar 17 2018

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/14b629ee440000
Cc: dtu@chromium.org
+dtu: looks like the test isn't producing values and there's not much we can do to bisect?

Comment 8 by dtu@chromium.org, Mar 19 2018

Yes, but that leads to the question: why did it produce values on the perf waterfall but not bisect?
Cc: simonhatch@chromium.org
I don't know! Simon, who do you think would lead an investigation of why a test produced values on perf dashboard but not pinpoint?
Think it should start with us making sure we're parsing the data correctly. There seems to be valid histogram output, Dave got time to look at this? Otherwise punt it to me and I'll dig through.

Comment 11 by dtu@chromium.org, Mar 19 2018

Looks like the story being passed in from dashboard is "nytimes", but the story in the histogram output is "http://nytimes.com/". So it can't find the story.

Comment 12 by dtu@chromium.org, Mar 19 2018

We've probably seen a bunch of these.
I think these were all migrated in  crbug.com/818035 

Might be able to bisect this if we specify the old name, hopefully it doesn't overlap the range where the names were changed.

Comment 14 by dtu@chromium.org, Mar 19 2018

Filed a tracking bug here: https://github.com/catapult-project/catapult/issues/4346

Comment 17 by dtu@chromium.org, Mar 19 2018

(Watch out for the webview bots, they're super overloaded)
Ah true, can you kill the job I started in #15 then?
Project Member

Comment 19 by 42576172...@developer.gserviceaccount.com, Mar 19 2018

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/158daeb9440000

Comment 20 by dtu@chromium.org, Mar 19 2018

Looks like it's not sufficient to just change the "story" in the bisect form. The "trace" comes from the code.
Project Member

Comment 21 by 42576172...@developer.gserviceaccount.com, Mar 22 2018

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/11924945440000
Project Member

Comment 24 by 42576172...@developer.gserviceaccount.com, Mar 28 2018

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/12dddceb440000
Dave, can you look at the pinpoint error?

Comment 26 by dtu@chromium.org, Mar 28 2018

Looks like the story changed name in dashboard, so the bisection was run on the wrong story name.
(https://github.com/catapult-project/catapult/issues/4346)
Project Member

Comment 27 by 42576172...@developer.gserviceaccount.com, Mar 28 2018

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/16f592ab440000
Status: WontFix (was: Untriaged)
Thanks, Dave!

WontFix-ing since this is a pretty small bump, and larger bumps later in the graph were able to be bisected, so not super worried about the story name change from #26.

Sign in to add a comment