Domain stuck on x-railway-fallback after a delete+recreate

Question

Looking for a help if anyone has
   seen this.
                                                                    
  The domain `api.getomaka.com` returns the Fastly fallback page    
  (404 "Application not found") even though:
                                                                    
  ✅ DNS resolves to the correct `requiredValue` (verified on       
  1.1.1.1 + 8.8.8.8)
  ✅ The customDomain is registered on the service per              
  `domains(...)` GraphQL query (status DNS_RECORD_STATUS_PROPAGATED 
  before I turned on Cloudflare proxy)
  ✅ Same service responds 200 on its other custom domain           
  (`api.ikanos.ro`) and on                                          
  `ecosystem-public-api-production.up.railway.app`
  ✅ TLS cert is valid (Let's Encrypt, fresh)                       
  ✅ Cloudflare proxy ON — same setup as the working domain       
  ✅ Service deployed twice successfully after the rebind
                                                                    
  Timeline:
  - ~21:57 UTC yesterday: I deleted the original Project that owned 
  this domain and let `ecosystem-public-api` (different project,    
  same team) pick it up.
  - ~07:09 UTC today: noticed Fastly was still routing to the old   
  per-domain CNAME. Did `customDomainDelete` + `customDomainCreate` 
  via API → got a fresh `requiredValue`.
  - ~07:10 UTC: updated Cloudflare CNAME to the new target.         
  - It's now 09:00+ UTC and still 404. `x-railway-fallback: true`.
                                                                    
  Looks like the host-based routing entry on Fastly never got       
  republished after the rebind. Anything I can do client-side to    
  force a refresh, or do I need someone with internal access to push
   it?                                                                                                                       
  Thanks!