All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=704491

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgzJnR4QoM


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

webrtc-win-large-tests

Stefan, do you see any reason why the congestion controller split would cause this? There might be some slight change in behaviour.

I also note that the graph went down a few weeks ago, see https://bugs.chromium.org/p/chromium/issues/detail?id=697089. 

My guess would be that it's because we're no longer calling OnRttUpdate on the receive-side cc, but only the send-side:

call.cc:347 call_stats_->RegisterStatsObserver(&send_side_cc_);

Seems like a bug.

I'll prepare a cl to address that.

See cl https://codereview.webrtc.org/2777423002/

The following revision refers to this bug:
  https://chromium.googlesource.com/external/webrtc.git/+/bcbaf746439291249b5688b239e6346caee90368

commit bcbaf746439291249b5688b239e6346caee90368
Author: nisse <nisse@webrtc.org>
Date: Tue Mar 28 08:16:25 2017

Let Call register ReceiveSideCongestionController as CallStatsObserver.

Fixes a regression from cl https://codereview.webrtc.org/2752233002.

BUG=chromium:704491,webrtc:6847

Review-Url: https://codereview.webrtc.org/2777423002
Cr-Commit-Position: refs/heads/master@{#17407}

[modify] https://crrev.com/bcbaf746439291249b5688b239e6346caee90368/webrtc/call/call.cc
[modify] https://crrev.com/bcbaf746439291249b5688b239e6346caee90368/webrtc/modules/congestion_controller/include/receive_side_congestion_controller.h

The supposed fix is now landed, but it seems to have made no change to these graphs. Not sure what to try next.