[CI] - 429 Too Many Requests

Required prerequisites

• Consult the security policy. If reporting a security vulnerability, do not report the bug using this form. Use the process described in the policy to report the issue.
• Make sure you've read the documentation. Your issue may be addressed there.
• Search the issue tracker to verify that this hasn't already been reported. +1 or comment there if it has.
• If possible, make a PR with a failing test to give us a starting point to work on!

Describe the bug

Problem

Our workflows make many requests to GHCR. This gets frequently rate limited.

See https://github.com/NVIDIA/cuda-quantum/actions/runs/22109358500/job/63901118985#step:11:314

#30 [prereqs 20/21] ADD .git/index /cuda-quantum/.git/index
#30 ERROR: failed to copy: httpReadSeeker: failed open: unexpected status code https://ghcr.io/v2/nvidia/buildcache-cuda-quantum-merge-group-assets-prereqs-cu12-6-llvm-arm64/blobs/sha256:f6d240177c52f64e99f6e65d17a5ccaa4e67a79a92a25789dd5d56c05da7e41b: 429 Too Many Requests - Server message: toomanyrequests: retry-after: 333.939671ms

Steps to reproduce the bug

Run multiple PRs/merge-queues at once.

Expected behavior

No rate limiting.

Is this a regression? If it is, put the last known working version (or commit) here.

Not a regression

Environment

CI

Suggestions

Maybe use multiple tokens, or different registries?

Start by inspecting the repository's CI workflows and the referenced GitHub Actions run, focusing on the GHCR requests that fail with HTTP 429. Determine how concurrent PR and merge-queue builds access the registry; done means the workflows avoid or appropriately handle the reported rate limiting when multiple runs execute.