All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=694620

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDguJzb9ggM


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

android-webview-nexus5X

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8987029255274202688

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Please try rerunning the bisect.


If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: android_webview_arm64_aosp_perf_bisect
  Benchmark    : blink_perf.layout
  Metric       : nested-percent-height-tables/nested-percent-height-tables

Revision             Result                  N
chromium@451449      691.821 +- 520.344      6      good
chromium@451666      44.3033 +- 13.9869      6      bad

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

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8987029255274202688

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5843967662358528


| 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 Speed>Bisection.  Thank you!

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8986483992401236912

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8986483968007574848

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Please try rerunning the bisect.


If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: android_webview_nexus6_aosp_perf_bisect
  Benchmark    : blink_perf.layout
  Metric       : nested-percent-height-tables/nested-percent-height-tables

Revision             Result                   N
chromium@451445      635.125 +- 7.05592       6      good
chromium@451663      34.3151 +- 0.554654      6      bad

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

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8986483992401236912

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5835529997778944


| 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 Speed>Bisection.  Thank you!

Finding a culprit CL...

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8985788751631203216

Bisects #6 failed due to infra issue. But visually looks like https://codereview.chromium.org/2701163002 made a problem.
Problem observed only on Android.

robhogan@ please take a look.
dominicc@ FYI.

That test was added with that very CL! But it was the second time it got added, so maybe the bots are confused.

First it landed here: https://codereview.chromium.org/2670603002 - r449101
Then that CL got reverted: https://codereview.chromium.org/2700183003 - r451451
Then an improved version landed: https://codereview.chromium.org/2701163002 - r451662

The graphs suggest that performance indeed went drastically down when that last CL landed, but previously the test was gone for a couple of days. Maybe the bots didn't run in that period, so they didn't notice that the test disappeared and then reappeared.

That said, maybe Rob wants to take a closer look at that test. The last CL fixed something to detect changes in table computed height (which the first (reverted) CL didn't care about). The test does exactly that: it changes the height of the containing block, which should affect the height of all tables. So maybe the test cannot be used anymore to show the performance improvements that this fix was supposed to give?

Strange that this is only seen in Android - the test is good, I see it regress to less than 1 run/s when I revert r451662 locally. Not really sure where to go with this.

If the test is good, I think you should just close this.

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Please try rerunning the bisect.


If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: android_webview_nexus6_aosp_perf_bisect
  Benchmark    : blink_perf.layout
  Metric       : nested-percent-height-tables/nested-percent-height-tables

Revision             Result                  N
chromium@451450      636.185 +- 6.95579      6      good
chromium@451663      34.552 +- 0.403175      6      bad

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

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8985788751631203216

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5879272058650624


| 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 Speed>Bisection.  Thank you!

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8982620224997420976