All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=824170

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=e31b9210831253add9139f063138b7f26edbeb1525290ff056f232c176239f91


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

chromium-rel-win10

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12e718d5440000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/12e718d5440000

Disable flaky ProxyBrowserTest.BasicAuthWSConnect on Windows Dbg and ASAN bots. by karandeepb@chromium.org
https://chromium.googlesource.com/chromium/src/+/20db52670d7515ae141cac7f2fef90894a29a910

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

I am not sure how disabling a test can cause a performance regression. Un-assigning myself.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/15a2490d440000

📍 Found significant differences after each of 15 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/15a2490d440000

DevTools: Refactoring node highlighting by illiam@google.com
https://chromium.googlesource.com/chromium/src/+/784ce4bc37ec5d29c0ff881c90b2746e3f621903

Flatten on v8::String::{Write*,Utf8Length} regardless of hint option. by yangguo@chromium.org
https://chromium.googlesource.com/v8/v8/+/e0ef8d05e1bd309436d83aad77bd4a189fa087f2

[turbofan] Add test for Number.{isFinite,isInteger} inlining by sigurds@chromium.org
https://chromium.googlesource.com/v8/v8/+/dfc2805ccf7e9fc6795191d0431ed87ba28e37d4

cros: Simplify LoginScreenClient by removing a wrapper. by jdufault@google.com
https://chromium.googlesource.com/chromium/src/+/a14bfd83d6b44ff1e83ea86a259a76f4df621038

[animation-worklet] Move work to a separate backing thread by petermayo@chromium.org
https://chromium.googlesource.com/chromium/src/+/d13bc9cfaa290602169197d0a2cf82fee1f016d7

Adding a |root_background_color| to RenderFrameMetadata by yiyix@chromium.org
https://chromium.googlesource.com/chromium/src/+/380375bf58399888d84eb1d8d4559c53dc3794bf

Ignore invalid MP3 packets from ffmpeg. by dalecurtis@chromium.org
https://chromium.googlesource.com/chromium/src/+/c7916b3e214429239af483abb78efe45e925c08b

Removed getViewsForPadding by amaralp@chromium.org
https://chromium.googlesource.com/chromium/src/+/68664907a937dbee846a874756778e5be21b857b

Avoid crashing in FPDFText_CountRects() for invalid start values. by thestig@chromium.org
https://pdfium.googlesource.com/pdfium/+/704a617ec011cda5c821d210c83f311f0ba5b15c

Roll src/third_party/pdfium/ 20c94774c..ef44194f4 (3 commits) by pdfium-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com
https://chromium.googlesource.com/chromium/src/+/efc7ec990e59cb3a9a5c7194e773d3171499eb87

Tidy up canvas code after ImageBuffer is removed by xlai@chromium.org
https://chromium.googlesource.com/chromium/src/+/7efe41bf9b0cf217759e48320b6e9b9b076249d2

Dashboard - Convert anomaly_config.GetAnomalyConfigDict to tasklet. by simonhatch@chromium.org
https://chromium.googlesource.com/catapult/+/83cb9dc1ae846b08a7d7a2953abc5b26f14ef1d3

[Dashboard] Add IS_SUMMARY diagnostic by eakuefner@chromium.org
https://chromium.googlesource.com/catapult/+/580f1ca2874ce1a0850c4dbfdcc3317e2a192e9b

Roll src/third_party/catapult/ 2a467ca51..734f737c6 (3 commits) by catapult-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com
https://chromium.googlesource.com/chromium/src/+/c27919454794b261493f721f2182d8a87338b97c

macOS Sandbox: Add timezone subpath to profile. by kerrnel@chromium.org
https://chromium.googlesource.com/chromium/src/+/3d6ecf141a0157cfd83844aef94630b2a6de8284

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

+yukishiino@ +haraken@ (benchmark owners), I'm not familiar with this metric. It looks too noisy (just because the pinpoint job found too many suspicious CLs, most of them look irrelevant). However, the first CL from the list in #6 looks like something that can change this metric. Could you please help illiam@ understand the metric and how to find out if his CL is at fault or not?

Thanks a lot!

Jeremy, would you take a look?

Jeremy is on vacation this week (I'm not sure how bad/urgent this bug is, just as an FYI)

Lucas, would you take a look and/or triage this issue?

I've looked at this and at the graphs around the regression, and there was a large change in the ref.

Looking at the pinpoint results, I think this is just noise and not a real regression.