[Gantt Chart] - 1.1 Benchmark current performance

Description

Before we start making speed changes, we need to figure out what's causing the bottleneck and get a definitive benchmark of current browser and server load time so we can figure out if we've improved.

Acceptance Criteria

There should not be any changes to code. If you do make a new branch for, after putting together the write-up, you should delete the branch rather than raising a PR.

• Browser Network tab timing (TTFB, content download, total time) recorded for the Gantt Chart page load, run at least 3 times to account for variance, using the stress-test seed data
• Server-side timer added around the entire Gantt data endpoint handler, logging total handler execution time to the console
• End-to-end load time measured and recorded against seed data using both of the above
• Short written summary (a few sentences, posted as a comment on this ticket or a linked doc) stating: current load time, how much of it is server-side vs. network, and whether this matches the previously reported 10-15 second load time

Proposed Solution

Browser: run local dev env, go to gantt chart page, and open dev tools Network tab. Refresh the page with Network tab open. Find the request going to the gantt endpoint, click on it, and look at the timing tab. Write down the TTFB (time to first byte), total time, and content download time. Repeat 2 more times and record all 3.
Server: add a startTime at the very top of the gantt data endpoint, and right before every return path, log the total time (Date.now() - startTime). Load gantt chart again in browser and write down logged time from console. Repeat 2 more times.

Mocks

No response

Open the Gantt Chart page with the stress-test seed data and use the browser Network tab to record TTFB, content download, and total time for the Gantt endpoint across three runs. Add temporary timing around the endpoint handler and compare its console timings with the browser measurements. Done means a short comment or linked document reports the measurements, server-versus-network split, and comparison with the 10–15 second claim; do not raise a PR.