Account: djmax1976@yahoo.com

Edge router ID: edge-500b32d4e9aaa4d3ff91d6ebb54285cd

Affected domains:

mystore.nuvanaapp.com → production frontend (service 25dd959e-488c-44a6-9cec-595223d3c734)

www.nuvanaapp.com → production frontend (same service)

api-staging.nuvanaapp.com → staging backend (service dbfaeb9d-838d-4301-a0b2-b4f2d90a7423)

Reference this diagnostic: edge returns 404 "Application not found" for all three, but requests don't appear in service logs; internal URLs work; DNS is correct; other domains on the same edge router work fine.

Problem

Three custom domains are attached, show syncStatus=ACTIVE in the API, have DNS_RECORD_STATUS_PROPAGATED, but return "Application not found" from the edge for every request:

1. mystore.nuvanaapp.com (production frontend, service 25dd959e-488c-44a6-9cec-595223d3c734) — just attached today
2. www.nuvanaapp.com (production frontend, same service) — attached April 23, blocking Let's Encrypt cert issuance for weeks
3. api-staging.nuvanaapp.com (staging backend, service dbfaeb9d-838d-4301-a0b2-b4f2d90a7423) — attached April 23, deleted + recreated today

In the same projects, these domains work correctly:

• api.nuvanaapp.com (staging backend, same service as api-staging) — attached April 23, has valid Let's Encrypt cert
• staging.nuvanaapp.com (staging frontend, service 3578cda7-946f-42f1-a649-dcda136c0693)

Diagnostic evidence

• Request to mystore.nuvanaapp.com: edge returns 404 "Application not found".
• Same request does NOT appear in the production frontend service's httpLogs — confirming the edge drops requests before reaching the service.
• Request to internal URL frontend-production-8d10e.up.railway.app reaches the service and returns 200. So the service is healthy and listening on the correct port.
• Even bypassing public DNS (curl --resolve) and hitting Railway's edge IP for the new CNAME target directly with the correct Host header returns 404. So it's not DNS, not propagation, not caching — it's the edge router not having the host→service mapping.

Already attempted

• Verified DNS via dig + nslookup; CNAME values match Railway's requiredValue
• Deleted and recreated each affected customDomain via API
• Created a serviceDomain for the production frontend (frontend-production-8d10e.up.railway.app); this internal URL works fine
• Redeployed the production frontend service (deployment a8378569-9f0c-462b-b760-8e71d7f978e0, status SUCCESS at 2026-05-18 12:11 UTC)
• Toggled project public/private setting

None of the above caused the edge router to start recognizing the affected hostnames.

Request

Please refresh the edge router mapping for these three domains so the edge stops returning "Application not found" and Let's Encrypt can complete its HTTP-01 validation for cert issuance.