All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=915855

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=0723b637e5235f38da8825aab04565d59d376456a9095ba5af7c932bf9c8da16


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

android-nexus5x-perf

system_health.memory_mobile - Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/1755e51e140000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/1755e51e140000

[android] Track signin-in and sync browsing activity. by msarda@chromium.org
https://chromium.googlesource.com/chromium/src/+/d5ad5b0a060c52821f67afc2989cc007f0723b6a
memory:chrome:all_processes:reported_by_chrome:skia:effective_size: 1.369e+05 → 2.213e+05 (+8.437e+04)

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/17f256da140000

CC+ perezju@ as he is listed as the owner for system_health.memory_mobile

My CL records additional UMA data about browsing activity being signed in to Chrome and syncing. I do not expect it to change the memory used by Chrome to load youtube or any other site. What I find suspicious is that this only affected the memory used to load youtube (I do not see any regression loading the other media websites in the tests). At this point I doubt this is related to my CL.

The pinpoint job is quite definitive, the change was caused by your CL, although:
- The magnitude is only 82KiB.
- Indeed it appears to affect only the single youtube story.

If I had to guess I think the reason is the metric was a bit noisy to begin with, and your CL just jiggled things enough to change the proportion in which we see different values.

Given that the magnitude of the regression is small, and affects a single story, I think it's fine to WontFix.

+ushesh FYI as something to think about on perf regression alerting, etc.

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/17f256da140000

[android] Track signin-in and sync browsing activity. by msarda@chromium.org
https://chromium.googlesource.com/chromium/src/+/d5ad5b0a060c52821f67afc2989cc007f0723b6a
memory:chrome:all_processes:reported_by_chrome:skia:effective_size: 1.37e+05 → 2.073e+05 (+7.029e+04)

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

 Issue 915854  has been merged into this issue.