Flaky rate limit when getting logs from GitHub

[huydhn]

For example https://hud.pytorch.org/pr/191508

API rate limit exceeded for user ID 1617424. If you reach out to GitHub Support for help, please include the request ID 921C:117C1F:7CD0663:1BA1C7A9:6A6A3B59 and timestamp 2026-07-29 17:41:45 UTC. For more on scraping GitHub and how it may affect your rights, please review our Terms of Service (https:\/\/docs.github.com\/en\/site-policy\/github-terms\/github-terms-of-service)

The token is probably from https://github.com/pytorch/test-infra/blob/main/aws/lambda/github-status-test/lambda_function.py#L16 (beed to confirm this). This has been a long time since we change anything on this lambda. I don't remember which token we use here. Once we find out, create an issue to GitHub to raise the limit

[huydhn]

I have just added a new token to github-status-test lambda to try to fix the rate limit issue there, i.e. https://ossci-raw-job-status.s3.amazonaws.com/log/90943003249. For the context, the lambda has a mechanism to rotate multiple GitHub tokens to avoid being rate limited. We are using 2 till date, one from pytorchupdatebot, one from I don't know whom (probably Suo). They are just regular fine-grained PAT with read permission to GHA. I'm adding a new one to increase the number to 3.

[huydhn]

#8493