Forum

Follow-up: this is a live data recovery incident.
I have now confirmed that /root/.n8n/binaryData still exists in the mounted n8n volume, but the PostgreSQL n8n database currently has no n8n application tables in public, and the n8n UI opens as a fresh workspace.
Please urgently check whether there are any internal snapshots, pre-repair volume states, temporary dumps, or AI repair artifacts that can be restored for my project.

Since Zeabur backups are retained for only 7 days, the remaining recovery window is now very limited and may expire within the next few days.
If any internal snapshot, temporary backup, pre-repair volume state, or AI-repair artifact still exists, I urgently request that it be preserved immediately before the retention window is lost.
Please prioritize this case as a time-sensitive data recovery request.

Hi,

Good news — all of your workflows and credentials are intact. There was no data loss. The issue was that your n8n service was pointed at the wrong PostgreSQL database, so it was opening an empty schema and showing you a fresh workspace.

What we found

• Your original PostgreSQL service (6979bafc7ff4377ae830fe17) has been running continuously for 58+ days with no restart and no data corruption.
• Inside that PostgreSQL instance there are two databases with n8n schemas:
• n8n — this is the one your DB_POSTGRESDB_DATABASE env var was pointing to. Tables exist but contain 0 workflows (likely created when you flipped the storage mode and n8n briefly initialized fresh tables here).
• zeabur — this contains all 38 of your workflows and 24 credentials, last updated 2026-04-02. Your data has been sitting safely here the whole time.
• Your n8n volume /root/.n8n/binaryData is also intact — about 525 MB of binary data.

What we did

We changed your DB_POSTGRESDB_DATABASE env var from n8n → zeabur and redeployed the n8n pod. After the restart, n8n loaded all 38 workflows from the correct database. You should now see your full workspace in the n8n UI.

One thing to do on your side

Some of your active workflows (e.g. 产品信息自动生成 - Google Sheet → Claude → 回填, AI图片生成工作流 - Google Sheets + 多模型-kie api) are failing to activate with OAuth errors:

The provided authorization grant (e.g., authorization code, resource owner credentials) or refresh token is invalid, expired, revoked, does not match the redirection URI used in the authorization request, or was issued to another client.

This is because the underlying Google / OAuth refresh tokens have expired during the downtime since 2026-04-02. You'll need to re-authenticate those OAuth credentials in the n8n Credentials UI (open each affected credential → click "Reconnect" / re-authorize). Once re-auth'd, the workflows will activate normally. The workflow definitions themselves are completely intact.

If you run into anything else, please reply here. Thanks for your patience!