The sfo identifier is not a valid region - the valid identifiers are us-west2, us-east4-eqdc4a, europe-west4-drams3a, and asia-southeast1-eqsg3a. You can override the stuck dashboard config by adding a railway.json (or railway.toml) to your repo with "deploy": { "multiRegionConfig": null }, since config as code always takes precedence over dashboard settings, which should unblock your deployments.

We've corrected this through a couple of different paths. The regions should resolve correctly now because of changes we added in the backend, but if they aren't resolving you can fix this by going into your service, and selecting Settings and scrolling down to Scale where you should see a drop down for setting the correct region now. That will let you correct this without having to contact support. This control was added today, and if the automatic recovery didn't happen, should let you get back up and running.