Try set custom start command in the Service Settings

Thanks. Before changing the start command, please note that the build completes successfully, but the failed deployment stops during CREATE_CONTAINER / container scheduling before the application process starts. There are no application runtime logs for the failed deployment, and the public /api/health endpoint still returns 404 Application not found.

The last known healthy deployment a6949940-8a3f-4d75-bc9e-6a57ed4ccf14, the service, database, and the READY persistent volume mounted at /data must be preserved.

Could you please inspect and clear any orphaned container or volume attachment and restore service-domain routing to the healthy deployment? If a custom start command is definitely required, please provide the exact command and the platform log/evidence indicating a start-command failure.

Follow-up: the public https://orderpilot-ops-production.up.railway.app/api/health endpoint is still returning 404 Application not found. Please confirm whether the container scheduling / service-domain routing issue is being investigated, and advise on the next safe action. We have not changed the start command or triggered another deployment.

Your service is deploying and serving normally again: the latest deployment is live and https://orderpilot-ops-production.up.railway.app/api/health is returning 200. Your data on the /data volume is intact and untouched. The earlier failures were a container issue on the host that has since been cleared, and your build was never the problem, so no start command change is needed. You're safe to deploy as normal from here.