Consider running benchmarks on "native connection" (i.e. no traffic shaping), to help get baselines

We should consider running something like a preliminary benchmark on "native connection" (i.e. no, or very little, traffic shaping/rate-limiting).

From https://docs.fastly.com/guides/debugging/debugging-with-webpagetest: "On the Test Settings tab under Advanced Settings, Connection should always be set to Native Connection during initial benchmarks."