All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=618667

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgvJ7lpwoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgvNW_qAoM


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

android-nexus9

===== BISECT JOB RESULTS =====
Status: completed


=== Bisection aborted ===
The bisect was aborted because The metric values for the initial "good" and "bad" revisions do not represent a clear regression.
Please contact the the team (see below) if you believe this is in error.

=== Warnings ===
The following warnings were raised by the bisect job:

 * Bisect failed to reproduce the regression with enough confidence.

===== TESTED REVISIONS =====
Revision         Mean    Std Dev   N  Good?
chromium@392889  19.801  0.739994  9  good
chromium@394047  19.613  1.12849   9  bad

Bisect job ran on: android_nexus9_perf_bisect
Bug ID: 618667

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests smoothness.tough_pinch_zoom_cases
Test Metric: mean_input_event_latency/mean_input_event_latency
Relative Change: 0.55%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus9_perf_bisect/builds/1818
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9010328057020475456


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5861677807960064

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!

Kicked off a bisect for just the blogger page which has a very clear regression

Moving this nonessential bug to the next milestone.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Kicked again: https://chromeperf.appspot.com/buildbucket_job_status/9009962969110207232

===== BISECT JOB RESULTS =====
Status: completed


=== Bisection aborted ===
The bisect was aborted because The metric values for the initial "good" and "bad" revisions do not represent a clear regression.
Please contact the the team (see below) if you believe this is in error.

=== Warnings ===
The following warnings were raised by the bisect job:

 * Bisect failed to reproduce the regression with enough confidence.

===== TESTED REVISIONS =====
Revision         Mean     Std Dev   N   Good?
chromium@393210  12.5511  0.803086  12  good
chromium@393274  12.2959  0.73808   18  bad

Bisect job ran on: android_nexus9_perf_bisect
Bug ID: 618667

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests smoothness.tough_pinch_zoom_cases
Test Metric: mean_input_event_latency/Blogger
Relative Change: 0.92%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus9_perf_bisect/builds/1824
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9009962969110207232


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5908052851556352

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!

Re-kicked bisect for the yahoo sports page which showed a clear step regression: https://chromeperf.appspot.com/buildbucket_job_status/9009948417820447312

===== BISECT JOB RESULTS =====
Status: completed


=== Bisection aborted ===
The bisect was aborted because The metric values for the initial "good" and "bad" revisions do not represent a clear regression.
Please contact the the team (see below) if you believe this is in error.

=== Warnings ===
The following warnings were raised by the bisect job:

 * Bisect failed to reproduce the regression with enough confidence.

===== TESTED REVISIONS =====
Revision         Mean     Std Dev   N   Good?
chromium@393210  19.263   0.87109   12  good
chromium@393339  19.1666  0.711217  18  bad

Bisect job ran on: android_nexus9_perf_bisect
Bug ID: 618667

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests smoothness.tough_pinch_zoom_cases
Test Metric: frame_times/http___sports.yahoo.com_
Relative Change: 0.68%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus9_perf_bisect/builds/1825
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9009948417820447312


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5887974147883008

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!

Kicked bisect for yahoo news page which has a tighter range for elevation: https://chromeperf.appspot.com/buildbucket_job_status/9009768256558641280

===== BISECT JOB RESULTS =====
Status: completed


=== Bisection aborted ===
The bisect was aborted because The metric values for the initial "good" and "bad" revisions do not represent a clear regression.
Please contact the the team (see below) if you believe this is in error.

=== Warnings ===
The following warnings were raised by the bisect job:

 * Bisect failed to reproduce the regression with enough confidence.

===== TESTED REVISIONS =====
Revision         Mean     Std Dev   N   Good?
chromium@393210  13.018   0.718351  12  good
chromium@393274  12.5312  1.01818   17  bad

Bisect job ran on: android_nexus9_perf_bisect
Bug ID: 618667

Test Command: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --also-run-disabled-tests smoothness.tough_pinch_zoom_cases
Test Metric: mean_input_event_latency/http___news.yahoo.com
Relative Change: 3.14%
Score: 0

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/android_nexus9_perf_bisect/builds/1828
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9009768256558641280


Not what you expected? We'll investigate and get back to you!
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5305621411266560

| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Tests>AutoBisect.  Thank you!

wjmaclean@ could you PTAL and let me know if you think https://codereview.chromium.org/1964133002 may have triggered this?  The bots aren't doing a good job at bisecting and looking at the change logs for the time where it jumped, your change looks like the only one that might have impacted it.

Is it possible that the calls that RenderView makes to any observers is adding overhead that wasn't there before (assuming the pinch zoom code even goes through the zoom level setting code)?

In webview_plugin.cc:

https://cs.chromium.org/chromium/src/components/plugins/renderer/webview_plugin.cc?rcl=1467026214&l=343

This is the only RenderViewObserver that overrides OnZoomLevelChanged, which is the observer message sent. This could be a bit expensive, as it looks at the frame tree to find the local roots.

But, the callsites you found in my CL aren't invoked for pinch-zoom at all, just for setting page zoom via the wrench menu (which Android doesn't use), so I'm guessing my CL isn't the real culprit here?

Thanks.  Back to the drawing board.

Closing - can't bisect and no suspicious CL's in the range