All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=637563

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICg9su9vwoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICg9sunoQoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICg9o-otwoM,agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICg9o-8rgoM


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

webrtc-android-tests-nexus5
webrtc-android-tests-nexus72

hlundin@: Do you have any idea about whether any of the listed culprit CLs could have caused this?

None of the CLs in the range should have any impact on the NetEq performance.

This hits only Nexus5 and Nexus72. There is also a slight regression for Nexus9, and there the blame range seems to point at 13712 specifically, which makes no sense at all (it's a name change for a file unrelated to the test).

kjellander@, any infra changes that could explain this?

No we haven't done any hardware changes for Android devices in a long time for these bots.
The only ones that I know of has been trybots only, which don't upload perf data.

Since the blamelist is:
https://chromium.googlesource.com/external/webrtc/+log/c4ac70078113472ed41d240efb299097442190ba..fb372f0074a0cb21fa36976316c0b5fe3b4ae7bc
is it not possible any of terelius@ changes could have caused this?

Sorry, my changes can't cause this. The files I modified are part of a stand alone tool, and not included in the webrtc-lib at all.

It seems we don't have anything that can explain this so I'll close it as WontFix then :(