We have a number of metrics that we currently gather (latency, jitter, loss, AEC, etc) during WebRTC sessions. In many cases, a mechanical examination of these metrics can reveal significant information about call quality degradation. We should come up with a scheme that continuously examines these metrics during a session, and alerts the user in some sensible fashion when they indicate some form of distress. Ideally, users would be able to drill down on these indicators to find out additional information, some of which might be actionable (e.g., by linking to relevant Sumo pages). I'm filing this as a meta bug, since it involves, at least: (a) Media code to detect session duress; (b) UX design; (c) UI implementation; and (d) Sumo work. There may be additional components identified during the media code work.
P1 for doing the breakdown
Assigning to Randell to do breakdown
Will become a meta, but it's not one until we do a bug breakdown.
Breakdown never happened, lowering priority.