Currently we have no way to determine if a change somewhere has had a negative impact on mobile web's performance until changes hit production. The result of this is that we firefight changes out of our control e.g. introduced by another extension
Let's work this quarter to upcoming prevent this from continuing to be a problem.
We can (and should) leverage the same solution that we will use for desktop: Start testing the code on the way to production. We will collect timing metrics and also number of requests per asset type and size. And then graph them in Grafana as a first step.