Right now we only have the Transfer Size metric for JavaScript, as part of the pageWeight metric. It's main purpose is to keep us aware of data bandwidth cost (and by extent, time required to transfer that data over different connections).
What's missing is information about what the "cost" was of that JavaScript in terms of user experience on the page. How long it took to initialise, whether it's a single uninterrupted block causing jank, or nicely spaced out. Whether it's doing many expensive DOM operations at load time, or whether it's making use of best practices like delegate event handlers and deferred computation.
There's a number of ways to capture this, and I propose we'll try multiple to see what works best.
This task is for capturing the information from the Long Task API, which gives us a high-level picture of any JavaScript execution blocks that took up more than a specified amount of time (e.g. 50ms).