Start with the deployment log

Identify whether the failure is in dependency installation, build execution, route creation, or runtime startup. The first failing step is usually more useful than the final summary error.

Common build problems

  • A dependency is referenced but not included in the lockfile.
  • The build command assumes a local binary or file that is not committed.
  • Environment variables required during build are defined only for runtime.
  • The selected Node.js version differs from local development.

Runtime failures

A deployment can build successfully and still fail once traffic arrives. Check request logs, environment variables, upstream availability, and route configuration before changing the application code.

Safe recovery

Use rollback when the previous deployment is known-good. Keep the failed deployment available for comparison instead of redeploying from a different commit during the incident.

Domain errors often look like app errors

Verify DNS and certificate state before changing application routing.

Troubleshoot domains and DNS