Hey Stephen, thanks for replying again.
We actually do that as well, but we learned that measuring that latency only gives you the time it took the response header's to reach the gateway, not the full payload. Here's what we currently have:
We noticed that the bigger the response, the higher the bufferingLatency, even when the targetLatency may remain fairly similar. Anyway, my 2 questions of the original post still stand
Best,
-Jaime