Your Postgres volume and data are intact. This is a known issue where the container image becomes stale after an incident. A normal redeploy reuses the same image, so it won't fix it. Open your Postgres service, press Cmd+K (or Ctrl+K) to open the command palette, and select "Redeploy source image" to pull a fresh image, which should resolve the connection issue.

It didn't work. This is not sustainable. I have tried restarting, redeploying, pulling source image. Still nothing works

Your Postgres service is actually running and healthy. The logs show regular checkpoint activity and a successful deployment, and the volume is in a READY state with your data intact. There is even a query from your application hitting the database (a constraint violation on storage_credentials), which confirms the database is accepting connections. The "Attempting to connect to the database" message is from the dashboard's Data tab UI, not the database itself. Try connecting to the database directly using the connection string from your service's Variables tab (e.g., via psql or a database client) to confirm it is reachable from outside the dashboard.

The problem is connecting through the UI. I want to speak with a human

Good news first: your database itself is healthy. The logs show normal checkpoint activity and your application is querying it fine, so your data and connectivity are not at risk. The issue is specifically the dashboard Data tab failing to open its own connection.

An engineer has picked this up and is looking at the Data tab on your service directly. In the meantime you can connect with an external client (TablePlus, psql, etc.) using the connection variables from the service's Variables tab. I'll update you here as we make progress.

It is still not working.

I have the same issue, let me know if u figured it out.

Quick update: I checked your Postgres directly and it is healthy and serving connections normally. Your applications are connected to it right now and reading and writing without any dropped connections, and the database is completing its regular internal checkpoints. Your data is intact.

What is failing is only the Data tab in the dashboard. It opens its own separate connection to display your tables, and that specific connection is what is breaking here, not Postgres. It is a known issue on our side that restarting and redeploying will not clear, and we are working on it.

Until the fix lands, you can reach all of your data by pointing any Postgres client such as TablePlus, DBeaver, or psql at the connection values in your service's Variables tab. That path is unaffected. We will follow up here once the Data tab itself is working again.