All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=811413

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=b41dd80d3664b5bafcd4614bbb6dfa070e9d3934591563a25f2fc5ae737b2793


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

android-webview-nexus5X

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/16dd2d2d840000

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/16dd2d2d840000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12f21b8d840000

📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/12f21b8d840000

[AFDO] Roll Android's profile to 66.0.3336.3_rc-r1 by gbiv@chromium.org
https://chromium.googlesource.com/chromium/src/+/d07b07609f7b538fdbb09f6c06ccba20218f2e9b

[CI] Make GraphicsLayer::client_ a reference by wangxianzhu@chromium.org
https://chromium.googlesource.com/chromium/src/+/3e9396db587c6ada6fd6815b2bb6aa2a429b0708

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

My CL just changed a pointer which is never null to a reference, which doesn't affect binary code except that some unnecessary null pointer checks are eliminated, so it should not affect performance.

Prior work: https://bugs.chromium.org/p/chromium/issues/detail?id=799629.

It was noted on said bug that blink benchmarks, from the perspective of AFDO, are heavily micro-benchmark-y. That said, I have some blink-related investigation to finish on that bug; I'll fold this into that.

Keeping this in its own bug because this was introduced by a new profile, not by turning AFDO as a whole on.

Automatic message: All alerts recovered.
Graphs: https://chromeperf.appspot.com/group_report?bug_id=811413

Can't see any obvious+persistent AFDO-related flakiness recently, so I'm closing this. If I missed something, tracking for general AFDO stability over time is in issue 849881.