All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=600486

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgqJaDvgoM


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

webrtc-linux-large-tests

r12203 == https://codereview.webrtc.org/1826093004, which fixes the alerts in  bug 597332 . :|

skvlad@, can you take a look? (I'm also cc'ing this week's perf sheriff)

This is the expected behavior. The original change caused regressions in some tests and unexpected improvements in others (such as https://bugs.chromium.org/p/chromium/issues/detail?id=597714) as it changed the timing of the first RTCP packet. The fix in r12203 restores the original timing and therefore reverts everything back to the state it was >1 week ago. If you zoom out on the graph you can see the original change that caused the unexpected improvement.

These tests measure the time it takes to ramp up to the target bandwidth utilization. Changing RTCP packet timing affects that time, but does not really result in an actual performance improvement.

 Issue 600489  has been merged into this issue.