All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=591005

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDAj8-zogoM


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

chromium-webrtc-rel-linux

I cannot see what possibly caused this. Assign to the expert on this.

The blame list contains a WebRTC roll: https://chromium.googlesource.com/external/webrtc/trunk/webrtc.git/+log/4b6381a..c422cbd

holmer, pbos: Can any of your CLs cause this?

Note that the regular expand rate changed too, but no alert was fired (for reasons unknown): https://chromeperf.appspot.com/report?sid=663e4ef1800006ac8aa738a519da8778f71617fabc97a1f91dede0a6556c6de2&start_rev=375497&end_rev=376190.

If this is an audio-only test I think https://chromium.googlesource.com/external/webrtc/trunk/webrtc.git/+/39535ca49e8a7e2154f7f0f40b8cfebb4b470da7 is a good candidate (preventing a busy-looping thread).

The name of the test is WebRtcPerfBrowserTest.MANUAL_RunsAudioVideoCall60SecsAndLogsInternalMetrics, so that indicates video too.

If the bitrate of the video was too low before then maybe that could have done it as well? Will close as an improvement.