All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=814216

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=4e2a385929a1690aa88fe1438f1b7fd308b4c6418db77f813e0654e0350b4a0e


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

android-nexus5
android-webview-nexus6

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/16b25327840000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14ba1422440000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14c4e692440000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/14c4e692440000

Renderer observer of frame submission by jonross@chromium.org
https://chromium.googlesource.com/chromium/src/+/a2ff4f82109df55045dee9f54985a98054f86dc4

Omit the parent LayoutView's scroll offset from FrameRect by szager@chromium.org
https://chromium.googlesource.com/chromium/src/+/cf91f7965c29e2bc93e81f79cef5d77481b00e6e

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

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/166e6e06440000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14cb244e440000

Looking at the graphs, it's clear the regression occurred after root-layer-scrolling was turned on. RLS did regress the graphs but this shows up as a regression earlier than the one linked with this bug. Those regressions are fixed (in this range: https://chromium.googlesource.com/chromium/src/+log/1c7a6ab994d205d84cab9cd153b16b0edec9efaf%5E..52c5105d39b7d838daabddf79f0aa4d225bfec27?pretty=fuller&n=1000 - several fixes from szager@ and chrishtr@) just before whatever caused this regression.

I'll try to get a bisect to point out a culprit since this seems like a pretty bad regression but I don't think we should block RLS on it.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/16cd6b8e440000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/14cb244e440000

Renderer observer of frame submission by jonross@chromium.org
https://chromium.googlesource.com/chromium/src/+/a2ff4f82109df55045dee9f54985a98054f86dc4

Omit the parent LayoutView's scroll offset from FrameRect by szager@chromium.org
https://chromium.googlesource.com/chromium/src/+/cf91f7965c29e2bc93e81f79cef5d77481b00e6e

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

There's two regressions in a small range. Some of the graphs only show one large regression because both patches fall into the same build. The regressions can be seen separately rather clearly in android-nexus5/thread_times.tough_compositor_cases / tasks_per_frame_browser: https://chromeperf.appspot.com/report?sid=17a475c5683524e935691ef96197befa1d8908eca1c6154bfd1dcef0d912fcf9&start_rev=534657&end_rev=539290

This alert is actually on jonross@'s CL so assigning to him to take a look. I'll file a separate bug for szager@'s regression.

I added rate limiting to my change identified in #12 in 540539. When you expand the graphs range you can see drop downs after that patch.

Since szager@'s patch is being tracked separately I'm closing this one out.

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/16cd6b8e440000

Renderer observer of frame submission by jonross@chromium.org
https://chromium.googlesource.com/chromium/src/+/a2ff4f82109df55045dee9f54985a98054f86dc4

Omit the parent LayoutView's scroll offset from FrameRect by szager@chromium.org
https://chromium.googlesource.com/chromium/src/+/cf91f7965c29e2bc93e81f79cef5d77481b00e6e

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

Ah, thanks Jon. Interestingly in the grouped graphs link above "tasks_per_frame_browser" didn't show the most recent runs. That graph has indeed recovered.

For completeness, the split out bug is  issue 821953