Measure performance

I need the page load and API timings for the new dashboard component measured end-to-end; run Lighthouse and the browser timing suite against the dashboard route, capture First Contentful Paint, Time to Interactive, and API RTTs for the authenticated endpoints, and hand me a short list of the three slowest requests to fix before Tuesday's demo.

Before I hand this to the performance team, make the measurements reviewer-friendly: put the key metrics (FCP, TTI, Largest Contentful Paint) at the top, include a short explanation of the test environment (Chrome stable, throttling: 4x CPU, 1x network), and highlight any requests or scripts that add more than 300ms to interactive time. Flag any single script over 50 KB parsed size.

Users reported the dashboard feels sluggish only on older laptops; I ran traces in staging and saw TTI jump from 1.2s to 3.8s on those devices. I don't know whether heavy parsing, main-thread work, or network congestion is the main culprit. Which diagnostic should I run next to pinpoint the dominant cause and what exact metric will show it?

Every release I find myself firefighting a single slowdown that slipped past review and cost a hotfix. I repeatedly wait until the demo to notice regressions. Where am I most likely failing in our review process and what single habit change would prevent these last-minute performance surprises?