About 8 hours ago, Railway automatically upgraded my services to Railway Metal, including my Postgres and Redis containers.

Immediately after the upgrade:

• My n8n instance went completely white

• I was logged out, and could no longer authenticate

• Logs showed repeated timeout exceeded when trying to connect errors from n8n trying to reach Postgres

I identified this as a Postgres connectivity issue caused by the upgrade, so I attempted to rollback the Postgres deployment (using the UI rollback button) to restore the previously working state.

However, the rollback has now failed multiple times with the error:

Pre-deploy › Migrate volumes

Volume migration failed

I also retried the rollback several times, and it fails at the pre-deploy step every time. Postgres cannot come back online, which leaves my entire n8n environment inoperable.

This Postgres instance contains production data (workflows, credentials, automations) for my n8n instance. I have not deleted or recreated anything, so I’m confident the data volume still exists.