All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=762206

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=36e92e089f1ac6286e79aad68b43f119ebf21955c522583142fc803976493178


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

android-nexus5

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8969256444006865248

=== Auto-CCing suspected CL author kylechar@chromium.org ===

Hi kylechar@chromium.org, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Perf regression found with culprit

Suspected Commit
  Author : kylechar
  Commit : 848a09e19e4417a1d5d3b6289e6716bfaf49086f
  Date   : Fri Aug 18 17:03:49 2017
  Subject: Flip default on FrameSinkManagerImpl.

Bisect Details
  Configuration: android_nexus5_perf_bisect
  Benchmark    : blink_perf.layout
  Metric       : nested-grid/nested-grid
  Change       : 5.38% | 15.2898693937 -> 14.4679233741

Revision             Result                   N
chromium@495542      15.2899 +- 0.528007      6      good
chromium@495567      15.178 +- 0.550627       6      good
chromium@495579      15.9978 +- 0.387324      6      good
chromium@495585      15.527 +- 0.364167       6      good
chromium@495586      15.6079 +- 0.270857      6      good
chromium@495587      14.6632 +- 0.107096      6      bad       <--
chromium@495588      14.7438 +- 0.233005      6      bad
chromium@495591      14.6678 +- 0.292396      6      bad
chromium@495640      14.4679 +- 0.272688      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests blink_perf.layout

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8969256444006865248


For feedback, file a bug with component Speed>Bisection

Will look into this. The CL was supposed to be a no-op, reversing a default value in the constructor and switching all instances of the constructor, but that doesn't appear to be the case.

The metric has improved again. I also wasn't able to reproduce locally or find any instances of FrameSinkManagerImpl that were changed by the suspect CL, so I'm unsure what the would have caused the regression.