Subject: Custom domain stuck in VALIDATING_OWNERSHIP 18+ hours; edge returns "Application not found"

Project: crisispr-api (03b948f1-8e87-4c37-839a-a29a671eeccb)

Environment: production (f144fdc1-b624-49f4-a72d-9d277d55e1de)

Service: api (15f98c9b-732a-4b96-97c6-c9c6a8702586)

Domain: api.crisispr.com.tw (custom domain id: 9815bd90-2f79-41c6-ab55-e67d9bfe7c33)

Timeline:

• Jun 7: customDomainCreate via GraphQL (id a8a2ae2b-...). DNS was added Jun 10 ~17:47 (+08).

• Stuck in VALIDATING_OWNERSHIP for 70+ min, so we deleted and recreated the domain

  (new id 9815bd90-..., Jun 10 19:10 +08). Cloudflare CNAME updated to the new target

  immediately (api -> x4umcs3v.up.railway.app, DNS only / grey cloud).

• 18+ hours later: certificateStatus still CERTIFICATE_STATUS_TYPE_VALIDATING_OWNERSHIP.

Evidence DNS is correct (your own API agrees):

• dnsRecords: requiredValue == currentValue == x4umcs3v.up.railway.app, status DNS_RECORD_STATUS_PROPAGATED
• dig @8.8.8.8 api.crisispr.com.tw CNAME -> x4umcs3v.up.railway.app (resolves globally)
• No CAA records on the zone.

Key symptom: curl -k https://api.crisispr.com.tw/healthz returns your edge 404

{"status":"error","code":404,"message":"Application not found"} and TLS serves the

default *.up.railway.app certificate. The hostname appears never activated at the edge,

so ACME validation can never succeed (chicken-and-egg).

Also possibly related: railway domain CLI mutations return "Unauthorized" for this

project even when logged in (while logged in to the correct account), while the same operations via GraphQL work.

Please kick/retry the certificate provisioning for this custom domain, or advise.

The service itself is healthy at https://api-production-523c.up.railway.app/healthz.