Space page and commit API both return 500 Internal Server Error

Beginners
1

Hi Hugging Face team,

My Space SAkizuki/DanbooruSearch is returning 500 Internal Server Error when I open its page (screenshot attached).

image
image1699×626 37.6 KB

Syncing from GitHub Actions using HfApi.upload_folder() also fails at the commit endpoint, even after all five automatic retries:

POST https://huggingface.co/api/spaces/SAkizuki/DanbooruSearch/commit/main
500 Internal Server Error

Request ID from the failed API call:

Root=1-6aaa4082-13a7e082564b8f3f1ed8ed5d;a9f32482-8952-4af0-9d3f-bad0293a606b

Could you please check whether this is a Hub issue or something specific to this Space, and advise on how to restore access and uploads?

Thanks!

2

Checked the Space itself, it’s actually up and running now (“Running on CPU Upgrade”), so whatever caused the 500 seems to have cleared on its own, that’s consistent with a transient Hub-side blip rather than anything specific to your repo (these generic “Internal Error - we’re working hard to fix this” 500s show up periodically across models/datasets/spaces and usually aren’t config issues). If the commit API is still failing when you try syncing again, worth double-checking it’s not just retrying into the same transient window, if it fails consistently on retry now that the page itself loads fine, that’d be worth a fresh report with the new request ID since the old one won’t be traceable anymore.

3

I’ve seen this exact issue before, and the symptoms strongly suggest a backend problem on the Hugging Face side rather than anything wrong with your Space.

A 500 Internal Server Error on both:

  • the Space page
  • the commit API endpoint

usually means the Space metadata or runtime state is stuck or corrupted on the Hub. This can happen even when the code hasn’t changed — especially after backend updates or infrastructure maintenance.

A few things you can check while HF staff looks into it:

  1. Try opening the Space’s “Files” view directly:
    SAkizuki/DanbooruSearch at main
    If this also returns 500, it confirms the issue is on the Hub side.

  2. Try restarting the Space from the Settings tab (if accessible).
    Sometimes the runtime gets stuck in a bad state.

  3. If the Space uses ZeroGPU or heavy model downloads, the startup logs may fail before the UI loads, but that still shouldn’t cause a 500 on the Space page itself.

  4. The fact that HfApi.upload_folder() fails at the commit endpoint with 500 — even after retries — is another strong indicator that the Hub backend is not accepting updates for this Space.

Given all this, it’s almost certainly an infrastructure issue. Posting the Space ID and the Request ID (which you already included) is the right move. HF staff can check the backend logs and clear the corrupted state.

Nothing in your description suggests a problem with your code or your GitHub sync pipeline.