Subject: Urgent: STARK production outage — container provisioning failures in AMS

Hello Railway Support,

Our production service stark in project energetic-courtesy is unavailable. Its public endpoint returns HTTP 502, and two deployment attempts failed during container creation.

Project: energetic-courtesy

Project ID: fbea822f-4bc8-45a6-b32e-583461e185c1

Service: stark

Service ID: da9f7f95-b78a-4256-bde1-af89fe8eb41e

Region: ams

Environment: production

Affected deployments:

• cfe56968-aa9f-431e-9011-9545342afeee — Failed October 21, 03:06 UTC
• e64b7ccc-094f-4595-8311-3571d97ddf85 — Failed October 21, 04:28 UTC
• 9183ad07-6c9e-4772-a313-383de4ed2cbb — Previously successful, still reported as active

Both failed deployments completed image building but produced no application runtime logs.

The previous successful deployment started correctly and listened on port 8080. However, Railway's proxy logs now show repeated HTTP 502 responses with connection refused, even though the instance is reported as running.

Critical data-preservation requirement:

The service has an existing persistent volume:

• Name: stark-volume
• ID: 3b49da21-7fb4-49cb-83fa-63ca8e9220df
• Mount: /app/data

Please do not delete, recreate, replace, or detach this volume.

Could your infrastructure team:

1. Investigate the container creation failures, including scheduling, image provisioning, and volume mounting in the AMS region.
2. Investigate why the previously successful instance is still reported online while its port refuses connections.
3. Restore the existing production service using its previous working deployment/image, preserving the attached volume and data.
4. Confirm the service is reachable and stable before closing the incident.

No application code, database, dependency, or infrastructure configuration changes have been made during this investigation.

Thank you for your assistance.

Eduard Gannaban