Label authenticated HTTP benchmarks as SSR transport evidence

When a full browser runner is unavailable, an authenticated cookie-preserving HTTP benchmark can measure server-rendered response timing, but it must be labeled SSR transport evidence rather than browser performance.

Use representative low, medium, and high-cardinality records; run at concurrency one; discard warmups; record time-to-first-byte and full-body samples; and require the expected authenticated status and route for every sample. Repeat the exact protocol after deployment.

This does not measure hydration, layout, client data fetching, or interactivity. Keep browser end-to-end coverage as a separate obligation and never log credentials or tokens.