Before I had assumed that it wasn't critical for these tests
to have the stats polled at a very regular interval but the perf
waterfall disagrees, so I'm accounting for drift when scheduling
the callbacks.
Bug: chromium:993688
Change-Id: If7f1d3919093f97508774c0c635fff6fe5081c10
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/149809
Commit-Queue: Tommi <tommi@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#28914}
6b117a5f3a