Troubleshooting

Build or start failed

Check:

• Start command is valid
• Required env vars are present
• Dependencies install successfully

Then redeploy after fixing config or code.

App is running but unreachable

Check:

• App listens on configured PORT
• Health endpoint (if used) returns success

For a custom domain, also confirm that the exact DNS records shown in the dashboard have propagated, the domain belongs to the active account, and verification has completed. The generated Autodisc domain can help separate an application-health problem from a DNS problem.

Missing env secrets

If runtime errors indicate missing keys/tokens:

• add env vars in dashboard
• redeploy

Do not paste secret values into support messages or the AI assistant. Share variable names and redacted logs.

Usage is exhausted

Open the usage page and confirm the active account, billing period, consumed GB-hours, and reset time. Paused compute can resume after an eligible upgrade or period reset. Exhaustion does not by itself delete project configuration or persistent data.

Connected login loops

Start from the Autodisc login page, then use Open Deploy again. Signing out from Autodisc Deploy clears both connected sessions. A managed account should not need a separate Dokploy-local password.