All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=635890

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICg1teXsAoM


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

chromium-webrtc-trunk-tot-rel-linux

The Chrome FYI bots found a regression. 

isheriff: I wonder if you know whether the CL https://codereview.chromium.org/2217623002 could have caused this?

Did you mean to paste a different CL url ?

Yes, it looks as if I pasted in the wrong URL. Sorry! Will take another look.

ehmaldonado@: Could the reason for this be the same as for https://bugs.chromium.org/p/chromium/issues/detail?id=637568

peah@: No, I don't think so.

brandtr@ I cannot make any sense of which CL could have caused this regression. Do you have any suggestion of what could have caused it?

When looking at the buildbot stdio, the only CL in the blamelist is this one: https://codereview.webrtc.org/2228683002. I'm not intimately familiar with this test, but this CL could plausibly be the reason for the change.

tkchin@: Would you mind having a look at this performance regression, to see if it is expected?

peah@: Do you think the 'agc_energy_diff_with_agc/speech_44kHz_16bit_stereo.wav_segment_1' change in the graphs is related to the 'Unique_frames_count/720p_H264' change?

No, that looks unrelated, and was probably incorrectly included in this issue by me. I'm looking at that in another regression issue.

I don't know enough about the test to comment.

But
https://codereview.webrtc.org/2228683002 has been updated by https://codereview.webrtc.org/2248883002/

which, if indeed was the cause for the regression, should go back to normal since the change is now iOS specific.

Yep, the follow-up CL has made the metrics go back up on Windows and Linux. The corresponding alert was apparently ignored in the dashboard though. Thanks for the info, tkchin@!

peah@: Not sure if you are tracking the "agc_energy_diff_with_agc" alert in another bug or not. If so, you could close this one.

Removing the agc_energy_diff_with_agc alert off this bug, and closing since the update in #12 fixed the regression.

Thanks! I'm tracking the agc issue in another bug.