Space stuck at "Starting" - health checker not responding - June 20 2026

Yeah. Hugging Face staff have been notified, so now we just have to wait.:sweat_smile:

Also, if you upload the same file to a space you created under a different name, it might work for the time being. It’s just a workaround, though…

For now, those asyncio-related messages are errors that can be safely ignored in recent versions of Gradio 6.x (the issue itself closed, but probably a recurrence), so there’s no need to worry. Maybe.

Workaround/Findings :

Related:

400 case:

Same symptom here — Docker Space hetchyy/quranic-universal-audio.

The container boots cleanly (gunicorn Listening at ``http://0.0.0.0:7860, all
workers up) and serves HTTP 200 on the public domain for /, /healthz, and
/livez. The image builds successfully (Pushing image → DONE). But the
runtime API stays at stage: APP_STARTING / BUILDING with
replicas.current: 0 / requested: 1 and never reaches RUNNING — the runtime
SHA is stale vs repo HEAD.

Tried: restart, factory reboot, pause→resume, a fresh no-op revision (built
fine, never promoted), and a hardware toggle (froze in BUILDING ~15 min after
the build had already completed). Nothing promotes a replica.

@hysts could you force-reconcile / reschedule this Space’s deployment if possible? Looks
like the same health-checker incident.

Same issue at Heretic Grimoire - a Hugging Face Space by heretic-org

Thanks for your patience, everyone. We heard back from our infra team, and the issue should now be resolved. Your Spaces should now load normally, but please let us know if any of them are still stuck on “Starting.”

Thanks for the update hysts. The issue is now resolved and my space is running normally again. Thanks again for your help.

My issue was resolved by @hysts ,it was
a HuggingFace infrastructure issue on
their side, not an app-side problem.

For context, here is what I observed:

  1. Container logs showed app running
    correctly on 0.0.0.0:7860
  2. Direct .hf.space URL worked fine
  3. But Space page stayed at “Starting”
    indefinitely
  4. Factory rebuild multiple times did
    not help

The fix came from HuggingFace team
internally — there was nothing to change
in the app code.

If your logs show the app running correctly
and your direct .hf.space URL works,
it is likely the same infrastructure issue.
I would suggest:

  1. Posting here tagging @hysts
  2. Submitting a support ticket at
    Enterprise Support – Hugging Face

Hope this helps! :folded_hands: