Private Docker Space stuck at “Build Queued” with no Docker build logs

Spaces
1

Hi, I’m having an issue with a private Docker Space that appears to be failing before the Docker build actually begins.

The Space is:

lnafi/DEX-Core-Staging

I’m deploying to the Space automatically from a private GitHub repository using GitHub Actions and Hugging Face Trusted Publishers/OIDC. The GitHub Actions deployment completes successfully, and the commit is received by the Hugging Face Space.

The Space also recognizes the Docker configuration in the root README.md. However, when Hugging Face attempts to build the Space, it enters a build error without showing any Dockerfile execution or useful build output.

The build log contains only:

===== Build Queued at 2026-09-15 15:34:01 / Commit SHA: 21c5b92 =====

I also installed the Hugging Face CLI locally, authenticated successfully as lnafi, and ran:

hf spaces logs lnafi/DEX-Core-Staging --build

The CLI returns only the same Build Queued line and no additional container/Docker build output.

I have already tried a Factory Reboot of the Space, but the behavior remains unchanged.

So far I have verified that:

  • GitHub Actions successfully uploads the files to the Space.
  • The Space receives the correct commit.
  • The Docker Space configuration in README.md is recognized.
  • Factory Reboot does not resolve the problem.
  • The build appears to fail before the first Dockerfile instruction is executed.
  • The dedicated CLI build-log command provides no additional logs.

The Space uses a custom Docker setup with a small FastAPI application, but since there is no Docker build output at all, this currently appears to be occurring before the application or Dockerfile is actually processed.

Could someone from Hugging Face check whether the builder/scheduler state for this Space is stuck or whether there is another backend issue affecting it?

Hugging Face username: lnafi
Space: lnafi/DEX-Core-Staging
Commit SHA shown in the build log: 21c5b92

I can provide screenshots or additional configuration details if useful.

Thanks!

2

This is a recurring platform-side glitch (there’s a whole thread from July with a dozen+ people hitting the exact same “no logs at all, factory reboot doesn’t help” pattern, sometimes it self-resolves, sometimes not, and it’s tied to the build worker/scheduler not picking up the job at all rather than anything in your Dockerfile). Given yours also comes in through Trusted Publishers/OIDC specifically, worth testing whether a manual commit (edit README directly on the Hub, or git push with a token instead of OIDC) triggers a normal build, that’d isolate whether it’s the OIDC publishing path itself getting stuck vs. a general builder queue issue affecting everyone. If a manual push builds fine, that’s useful info to include when you escalate to website@huggingface.co, since it’d point at the Trusted Publishers integration specifically rather than a platform-wide outage.