fix: add error handling for speed test failures

[Aaradhya-07]

The startTest() method in app.js awaits runNdt7() and runMSAK() without any error handling. If either test fails — due to a network error, server timeout, or any thrown exception — the UI remains stuck in the "Testing" state with no way for the user to retry. This PR wraps the test execution in a try/catch/finally block: on success, results display as before; on failure, the user sees a "Test failed. Please try again." message; and in both cases, the finally block resets testRunning and updates the UI so the start button becomes clickable again.

[robertodauria]

Thanks for the contribution. The idea of handling test failures gracefully is sound - if runNdt7() or runMSAK() throws, the UI does get stuck. However, there are a few issues with the implementation:

• The error UX needs more thought: what happens to the gauge? Are partial results from the first test still displayed? Should there be an explicit retry affordance?
• runPT() is inside the try block but not awaited, so its errors won't be caught.

This also targets sandbox-roberto-rewrite, which is an active rewrite tracked in #83. Once that lands on main, it would be a good time to revisit this with a more complete approach to error states. Feel free to re-open against main at that point, ideally after coordinating with a maintainer on the UX.

On a broader note: we've been receiving a large volume of unsolicited, sometimes AI-generated PRs with no prior discussion or contact, and as a small team this is not sustainable. We are updating our contribution requirements. Going forward, please coordinate with a maintainer before opening PRs or issues. See https://github.com/m-lab/gsoc#2-engage-with-us for details.