All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=814149

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=9a6cae850af5b4cd68047e93717bb9c08edfef25cdcb70cf12fa9a9cf05bb080


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

android-nexus5X

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/11bc83fb840000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/11bc83fb840000

Revert "Add hooks for AXTreeDelegate when a relation target changes" by caseq@chromium.org
https://chromium.googlesource.com/chromium/src/+/a9e3124b7b00bcce07edf062b27dc107d5f4b401

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

I don't think this makes sense. First, this is just a revert of a change, that has been re-landed since (https://chromium-review.googlesource.com/c/chromium/src/+/924769). Second, the revert had to be done because the original change broke telemetry unittests. Not sure if perf results can be trusted in case telemetry is broken.

Alex, it's not quite clear to me what your expectations were when you assigned this to me.

Ah, sorry, just realized the assignment happened automatically by bisect bot -- apologies to Alex for implying he did that :-)

The graph reverted back to the previous median so LGTM on WontFix.