All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=747898

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=40e438c4b00aa92b23e610195bfb2de9058e5f4c2756bab972508bb935c520fa


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

android-nexus6

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8973191354912188016

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

Hi ntfschr@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 : Nate Fischer
  Commit : 94bc1feb17c446042f8b0e55807c5433ca8599b6
  Date   : Thu Jul 20 19:37:14 2017
  Subject: AW: fetch Safe Browsing user opt-out earlier

Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:proportional_resident_size_avg/load_search/load_search_google
  Change       : 0.55% | 105936116.889 -> 106353681.333

Revision             Result                   N
chromium@488346      105936117 +- 798377      9       good
chromium@488355      106145614 +- 837459      14      good
chromium@488356      106306480 +- 430214      21      bad       <--
chromium@488357      106363214 +- 484673      14      bad
chromium@488359      106350951 +- 573005      21      bad
chromium@488364      106320865 +- 565380      14      bad
chromium@488382      106133863 +- 602940      9       bad
chromium@488417      106223064 +- 631543      9       bad
chromium@488487      106353681 +- 328207      6       bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 --story-filter=load.search.google system_health.memory_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8973191354912188016


For feedback, file a bug with component Speed>Bisection

I'm going to put the status to "Assigned" please pass back to me and put the status to "Untriaged" if you think this is the wrong CL. Thanks!

Upon first inspection, this doesn't seem to be my CL. The benchmark uses "--browser=android-chromium" but my CL only touches android_webview/ folder, not Chrome. On top of that, my CL simply calls into closed-source code via reflection, which I wouldn't expect to increase memory.

I'm setting the status to untriaged, but if you think my analysis is wrong, please assign back and I can run benchmarks locally and continue investigation.

Hmm, yeah that doesn't seem likely. I'll click off a new bisect.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8972991694172438960

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

Hi thestig@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 : Lei Zhang
  Commit : 1909c58263b12c60b36b7ad54882805eded9daa8
  Date   : Thu Jul 20 19:19:49 2017
  Subject: PDF: Do not report metrics for print preview.

Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:proportional_resident_size_avg/load_search/load_search_google
  Change       : 0.63% | 106272444.0 -> 106918417.333

Revision             Result                   N
chromium@488346      106272444 +- 211619      9       good
chromium@488347      106477244 +- 208293      6       bad       <--
chromium@488348      106519228 +- 261668      6       bad
chromium@488349      106485436 +- 235079      6       bad
chromium@488351      106494652 +- 317017      6       bad
chromium@488355      106492019 +- 536096      14      bad
chromium@488364      106491466 +- 361721      9       bad
chromium@488382      106415121 +- 150953      6       bad
chromium@488417      106595516 +- 355128      6       bad
chromium@488487      106918417 +- 273116      6       bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 --story-filter=load.search.google system_health.memory_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8972991694172438960


For feedback, file a bug with component Speed>Bisection

Sorry, but it's not me either. The code in pdf/ isn't used on Android.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8972190220013302528

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

Hi droger@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 : David Roger
  Commit : a7dbb8484b076a69bf3d96ca34314a6f75f50e2b
  Date   : Thu Jul 20 23:00:35 2017
  Subject: Cleanup namespaces and includes in scoped_feature_list

Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:chrome:all_processes:reported_by_os:system_memory:proportional_resident_size_avg/load_search/load_search_taobao
  Change       : 0.48% | 110548973.333 -> 111077052.0

Revision             Result                    N
chromium@488000      110548973 +- 486633       6       good
chromium@488250      110742887 +- 503500       9       good
chromium@488375      110786017 +- 703926       14      good
chromium@488435      110814225 +- 835341       9       good
chromium@488452      110956951 +- 1007687      14      good
chromium@488460      110732769 +- 1039367      14      good
chromium@488462      110732419 +- 489299       9       good
chromium@488463      110706071 +- 770799       14      good
chromium@488464      111025413 +- 992519       14      bad       <--
chromium@488468      111335783 +- 722105       9       bad
chromium@488500      111143246 +- 582660       14      bad
chromium@489000      111077052 +- 273433       6       bad

Please refer to the following doc on diagnosing memory regressions:
  https://chromium.googlesource.com/chromium/src/+/master/docs/memory-infra/memory_benchmarks.md

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 --story-filter=load.search.taobao system_health.memory_mobile

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

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8972190220013302528


For feedback, file a bug with component Speed>Bisection

Nope. Not this one either. Looks like bisecting is unreliable unfortunately.

I think this metric is too noisy to bisect :(