Automatic deployments stopped triggering entirely a few days ago, right around the Sep 18 2026 "deployments stuck in initializing state" incident. I've been through Vercel support already (Hobby plan, so email/case support isn't available — posting here as suggested).
What I've tried, in order:
- Found the project had somehow gotten connected to the wrong repo (a different project of mine, vergentia) instead of biosync-app. Disconnected and reconnected to the correct repo, niwdecr-cell/biosync-app — confirmed correct in Settings > Git.
- Verified my GitHub account email matches my Vercel login email.
- Pushed a fresh empty commit after reconnecting. No deployment triggered.
- Checked GitHub > Settings > Webhooks for this repo: zero webhooks from Vercel, not even a failed one — even after the fresh reconnect. (I understand Vercel uses the GitHub App at the account level rather than per-repo webhooks, so this part may be expected — flagging for context.)
- Tried a manual "Redeploy" from the dashboard — it succeeded, but rebuilt an OLD commit, several behind my actual latest on main.
- Created a new Deploy Hook (branch: main) and POSTed to it directly. Got a 201 Created with a PENDING job ID (WoqSkSTrw7OBLX2X2Dmo). Waited several minutes — no deployment ever appeared in the Deployments tab.
So: the Git connection looks correct in the UI, manual redeploy works but on stale data, and even a direct Deploy Hook API call that returns success produces nothing. This matches the pattern in a few other threads here describing a stale server-side installation/build-dispatch state that survives a normal disconnect/reconnect after the Sep 18 incident.
Has anyone found a way to clear this from the user side, or does this need a Vercel team member to look at the backend state directly? Happy to provide more details (commit SHAs, job IDs, timestamps) if useful.
This has now blocked all production deployments for several days — any pointers appreciated.