Investigate improving Lighthouse scores for the JavaScript SDK

[HazAT]

Use the Lighthouse dashboard and the merged CI work as the starting point for investigating how to improve the performance impact of the JavaScript SDK across the tested configurations.

The goal is to understand where the SDK is affecting Lighthouse metrics, especially Largest Contentful Paint, and identify changes that could improve those scores over time.

Starting points:

• Dashboard: Lighthouse dashboard
• CI groundwork: ci: Add Lighthouse Test as Nightly CI run #20850

This should sit in the backlog until the team is ready to pick it up.

[linear-code]

JS-2589

[Lms24]

We've identified #21817 as a high contributor for overhead in default Lighthouse (Lantern mode) conditions. Gonna fix this next week.

[Lms24]

Closing this since at this time, we've investigated and identified the largest contributor to bad lighthouse scores (fixed via #21817). There are definitely things we can do to further improve initial SDK performance (lazy loading, mostly) but this needs to be investigated and experimented with in a larger project (after v11). We'll create new issues for this once we get to it.