This is a persistent platform issue. The Dockerfile revert didn't help because the problem isn't the code—it's the container scheduler. The diagnosis shows:

• The image builds and publishes successfully
• The container creation fails every time
• The old May 6 deployment still runs fine
• Memory usage is low (~87 MB)
• Even the exact same commit that succeeded on May 6 now fails

You need to contact Railway support. This is a platform-level infrastructure issue preventing new containers from being scheduled for this service. The support team can investigate why the container scheduler is failing.

In the meantime, your old deployment from May 6 is still running and healthy, so the service is operational—you just can't deploy new versions until this is resolved.