Hitting the same issue as https://station.railway.com/questions/custom-domain-stuck-in-certificate-statu-4835829b — likely a cdnMode desync.

Setup

- Project: matrixfood (production)

- Service ID: fda2fcc3-f91f-46c4-b5e9-9e2152fa599b

- Custom domains: matrixfood.com.br (apex) and www.matrixfood.com.br

- DNS managed by Cloudflare, DNS only mode (grey cloud)

Symptoms

- Both custom domains return HTTP 502 Bad Gateway with header X-Railway-Fallback: true for 24+ hours

- Railway-provided domain (matrixfood-production.up.railway.app) returns HTTP 200 OK

- SSL certs successfully issued by Let's Encrypt for both domains (verified via openssl)

- DNS records correct in Cloudflare (CNAME apex → 0yqdzr9o.up.railway.app, CNAME www → abm7mcgi.up.railway.app, plus required _railway-verify TXT records)

History

These domains were previously on another Railway account (edisonpncode/matrix-food). Removed from there, then added here. The Railway dashboard intermittently shows a stale "23 changes to apply" prompt with values that don't match the actual service state.

What I tried

- railway restart -s matrixfood

- railway redeploy -s matrixfood (full rebuild) — SUCCESS, still 502

- Latest deploy: 51716a8c-cbb7-4d27-983d-ca84c055ec2c

Could a Railway engineer check the cdnMode state for these 2 domains and re-sync them with the Fastly edge? Or alternatively delete the domain records server-side so I can re-add cleanly? Thanks.