Filmstrip and other assets throw 500s errors intermittently

[sergeychernyshev]

Filmstrip and other assets throw 500s errors intermittently (see screenshot)

[Judyzc]

Commenting for documentation later, I'm checking logs on staging environment with the command npx wrangler tail --env staging --format pretty.

I get a series of responses like this:

 GET https://staging.telescopetest.io/api/tests/REDACTED/filmstrip/frame_1366x768_1.jpg - Exception Thrown @ 3/24/2026, 12:09:35 PM
  (log) [R2] Request started - testId: 2026_02_20_14_43_51_2a6e06c6-6b81-4522-9f83-c324f720ae54, file: filmstrip/frame_1366x768_1.jpg
✘ [ERROR]   Error: The Workers runtime canceled this request because it detected that your Worker's code had hung and would never generate a response. Refer to: https://developers.cloudflare.com/workers/observability/errors/


GET https://staging.telescopetest.io/api/tests/REDACTED/REDACTED.webm - Ok @ 3/24/2026, 12:09:35 PM
  (log) [R2] Request started - testId: 2026_02_20_14_43_51_2a6e06c6-6b81-4522-9f83-c324f720ae54, file: 9b74e9e53ffc1584885637faa2eb116d.webm
  (warn) Warning: A promise was resolved or rejected from a different request context than the one it was created in. However, the creating request has already been completed or canceled. Continuations for that request are unlikely to run safely and have been canceled. If this behavior breaks your worker, consider setting the `no_handle_cross_request_promise_resolution` compatibility flag for your worker.
    at Object.resolve (<anonymous>)
    at index.js:41676:64

Here's the documentation from Cloudflare workers for why workers runtime is canceling: https://developers.cloudflare.com/workers/observability/errors/#cause-1-unresolved-promises. I think Prisma is getting connection pool exhaustion, each /result/{testId} load calls ~37 (? this count is from AI) requests from Prisma.

[sergeychernyshev]

The problem is reduced with #224 but we need to work on #209 to remove rating check from serving the assets.