Audit: Total count of RTT to FCP/FMP/etc

@paulirish and I discussed about surfacing the # of round-trips to FMP in Lighthouse / DevTools. The advantage of this metric is that it's directly controlled by the developer and isn't dependent on throttling settings. For example, conditions such as TCP slow start, TLS handshake, latency & bandwidth constraints are for all intents and purposes out of the control of developers.

Today, Lighthouse surfaces the critical request chain based on network priority (I'm not too familiar with this so correct me if I'm wrong) and the duration until FMP.

Two ideas:
* 1) short-term - show # of round trips till FMP in the Lighthouse report (just talked with @patrickhulce and he's already doing something along these lines!)
* 2) longer-term - prototype a "perf" mode in Network panel of DevTools that highlights the critical rendering chain and displays the # of round trip till FMP in the summary bar.

Start by reviewing Lighthouse’s existing critical request chain and FMP reporting, then examine the proposed Network panel “perf” mode in DevTools. The scope must first be decided between surfacing round trips in the Lighthouse report and prototyping the DevTools experience; done means the selected proposal has a defined implementation and validation path.

Written by the indexing model from the issue text.