All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=747485

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=5ef953be277a39fc296d61874d2f17b34b408ceb40e2d4e7ce584b1f03c1027f


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

android-nexus5
android-nexus6
android-nexus7v2
android-webview-nexus5X
android-webview-nexus6

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8973437917996973264

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

Hi hubbe@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 : hubbe
  Commit : e2cc88c09afab5f066c83032548c1180d99c5ae7
  Date   : Fri Jul 14 23:11:01 2017
  Subject: media: Allow suspend on HTTP 200 videos

Bisect Details
  Configuration: android_webview_nexus6_aosp_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:webview:all_processes:reported_by_os:system_memory:java_heap:proportional_resident_size_avg/browse_media/browse_media_imgur
  Change       : 10.53% | 5321159.11111 -> 5881287.11111

Revision             Result                   N
chromium@486073      5321159 +- 11285356      9      good
chromium@486875      4340053 +- 134816        6      good
chromium@486901      4746240 +- 277245        6      good
chromium@486914      4916907 +- 126568        6      good
chromium@486917      5161301 +- 465556        6      good
chromium@486918      5437099 +- 253583        6      good
chromium@486919      6336853 +- 325492        6      bad       <--
chromium@486920      6241280 +- 416414        6      bad
chromium@486926      5634389 +- 146505        6      bad
chromium@486976      5437099 +- 514873        6      bad
chromium@487076      5394091 +- 224472        6      bad
chromium@487276      5364053 +- 506835        6      bad
chromium@487676      5881287 +- 730795        9      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-webview --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=browse.media.imgur 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/8973437917996973264


For feedback, file a bug with component Speed>Bisection

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8973093290627239616

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8973092699738494832

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8973092680068961744

There are some ~0.9 to 2.0mb (5% - 7%) increases in renderer malloc for various browse stories in the same range (see [1]). Think they might have the same cause,
bisecting to be sure.


[1]: https://chromeperf.appspot.com/report?sid=579369d1114b064b3d217322157df09efb9108fb2ec4ccc8b08445fa115faf4a&start_rev=473906&end_rev=489266&rev=488787

=== BISECT JOB RESULTS ===
Perf regression found but unable to narrow commit range

Build failures prevented the bisect from narrowing the range further.


Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:malloc:effective_size_avg/browse_social/browse_social_instagram
  Change       : 3.97% | 14911966.2222 -> 15503245.3333

Suspected Commit Range
  12 commits in range
  https://chromium.googlesource.com/chromium/src/+log/4afd513b88234976f17b06eee24892ce9d149e7e..e5ef88bfc930b8066ae7c8a5ec5cf35577bbcd2b


Revision                           Result                   N
chromium@486073                    14911966 +- 1079311      9        good
chromium@486474                    14625156 +- 769698       6        good
chromium@486675                    14952192 +- 360710       6        good
chromium@486725                    14569028 +- 531590       6        good
chromium@486738                    14744473 +- 1064136      9        good
chromium@486744                    14742782 +- 867046       9        good
chromium@486745                    14648569 +- 958846       6        good
chromium@486745,v8@44cce0fcc0      ---                      ---      build failure
---                                ---                      ---      too many build failures to list
chromium@486745,v8@a97abb909b      ---                      ---      build failure
chromium@486746                    15989705 +- 761471       6        bad
chromium@486747                    15772589 +- 1189765      6        bad
chromium@486750                    15955316 +- 1168916      6        bad
chromium@486775                    15847440 +- 490174       6        bad
chromium@486875                    15772443 +- 951563       6        bad
chromium@487676                    15503245 +- 1037220      9        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=browse.social.instagram 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/8973093290627239616


For feedback, file a bug with component Speed>Bisection

=== BISECT JOB RESULTS ===
Perf regression found but unable to narrow commit range

Build failures prevented the bisect from narrowing the range further.


Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:malloc:effective_size_avg/browse_media/browse_media_flickr_infinite_scroll
  Change       : 7.80% | 19757832.0 -> 21299212.0

Suspected Commit Range
  12 commits in range
  https://chromium.googlesource.com/chromium/src/+log/4afd513b88234976f17b06eee24892ce9d149e7e..e5ef88bfc930b8066ae7c8a5ec5cf35577bbcd2b


Revision                           Result                   N
chromium@486073                    19757832 +- 856514       6        good
chromium@486474                    19861007 +- 931310       6        good
chromium@486675                    20019827 +- 520621       6        good
chromium@486725                    20147527 +- 403418       6        good
chromium@486738                    19996017 +- 423346       6        good
chromium@486744                    20219407 +- 577876       6        good
chromium@486745                    19979601 +- 473489       6        good
chromium@486745,v8@44cce0fcc0      ---                      ---      build failure
---                                ---                      ---      too many build failures to list
chromium@486745,v8@a97abb909b      ---                      ---      build failure
chromium@486746                    21667337 +- 817240       6        bad
chromium@486747                    21392265 +- 827042       6        bad
chromium@486750                    21476360 +- 375980       6        bad
chromium@486775                    21292004 +- 893693       6        bad
chromium@486875                    21780344 +- 519033       6        bad
chromium@487676                    21299212 +- 1079221      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=browse.media.flickr.infinite.scroll 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/8973092699738494832


For feedback, file a bug with component Speed>Bisection

=== BISECT JOB RESULTS ===
Perf regression found but unable to narrow commit range

Build failures prevented the bisect from narrowing the range further.


Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : system_health.memory_mobile
  Metric       : memory:chrome:renderer_processes:reported_by_chrome:malloc:effective_size_avg/browse_social/browse_social_pinterest_infinite_scroll
  Change       : 4.09% | 26685509.3333 -> 27776692.0

Suspected Commit Range
  12 commits in range
  https://chromium.googlesource.com/chromium/src/+log/4afd513b88234976f17b06eee24892ce9d149e7e..e5ef88bfc930b8066ae7c8a5ec5cf35577bbcd2b


Revision                           Result                   N
chromium@486073                    26685509 +- 591332       6        good
chromium@486474                    26780196 +- 481435       6        good
chromium@486675                    26729804 +- 879756       6        good
chromium@486725                    26827513 +- 728764       6        good
chromium@486738                    26648836 +- 566383       6        good
chromium@486744                    26796772 +- 151686       6        good
chromium@486745                    26321207 +- 831888       6        good
chromium@486745,v8@44cce0fcc0      ---                      ---      build failure
---                                ---                      ---      too many build failures to list
chromium@486745,v8@a97abb909b      ---                      ---      build failure
chromium@486746                    28006827 +- 1263534      6        bad
chromium@486747                    28005187 +- 442922       6        bad
chromium@486750                    27719465 +- 717675       6        bad
chromium@486775                    27816825 +- 745097       6        bad
chromium@486875                    27721880 +- 705059       6        bad
chromium@487676                    27776692 +- 408198       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=browse.social.pinterest.infinite.scroll 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/8973092680068961744


For feedback, file a bug with component Speed>Bisection

I suspect this is a symptom of caching the whole video when the video is small, which is intended.

What's the limit of the intended regression?

Only videos 25Mb or smaller are fully cached.
The regression should be smaller than that, since we would previously allow the cache to grow up to 25Mb, but didn't care if some parts got thrown out. After my change we explicitly keep all parts of the video if we think it's small enough.