Skip to content
Apex haidaratelier.com served Railway’s wildcard certificate while www worked
haidaratelier-star
OP

14 minutes ago

Could Railway check the TLS certificate order for haidaratelier.com on service haidar-atelier-v6 (production, port 8080, Hobby) and tell me how to retry it safely?

On 19 November 2026 at about 02:47–02:57 UTC, the apex used a Namecheap ALIAS to nv32sez8.up.railway.app. The authoritative nameserver and public resolvers returned only 153.52.255.124, and the Railway ownership TXT matched. There was no conflicting A/AAAA answer or restrictive CAA record. Plain HTTP reached Railway and redirected to HTTPS, but a strict TLS connection with SNI haidaratelier.com failed with ERR_TLS_CERT_ALTNAME_INVALID. The certificate listed only *.up.railway.app and up.railway.app.

The www.haidaratelier.com binding on the same service (k2v30x62.up.railway.app) served a valid certificate and returned a 308 to the apex. The Networking panel showed “Cloudflare proxy detected” for both hosts. Namecheap hosts our DNS, and the test route pointed directly to Railway. I did not see a certificate retry button.

The original website routing has been restored to keep the site online. Both Railway domain bindings and verification TXT records remain in place. Could you check whether the apex certificate order was created, failed, or is still pending, and whether the proxy badge is related? What is the supported way to retry issuance and verify it before committing to the cutover? I can provide project and service IDs privately if needed.

Awaiting User Response

3 Replies

Railway
BOT

13 minutes ago

The apex domain does not have a certificate yet. It is still at the ownership-validation stage, which comes before a certificate is issued. The *.up.railway.app certificate you saw is what our edge serves for a name that has no certificate of its own, and www works because it already has one.

Your ownership TXT record is verified. Before we place a certificate order, we also check that plain HTTP requests for the domain reach us. Our most recent check, at 06:14 UTC today, found the apex was not being routed to us. That matches the original routing you restored after your test.

Issuance can only go ahead while the apex points at your target (nv32sez8.up.railway.app) and HTTP requests to it reach us.


Status changed to Awaiting User Response Railway • 13 minutes ago


Railway

The apex domain does not have a certificate yet. It is still at the ownership-validation stage, which comes before a certificate is issued. The `*.up.railway.app` certificate you saw is what our edge serves for a name that has no certificate of its own, and `www` works because it already has one. Your ownership TXT record is verified. Before we place a certificate order, we also check that plain HTTP requests for the domain reach us. Our most recent check, at 06:14 UTC today, found the apex was not being routed to us. That matches the original routing you restored after your test. Issuance can only go ahead while the apex points at your target (`nv32sez8.up.railway.app`) and HTTP requests to it reach us.

haidaratelier-star
OP

12 minutes ago

Thanks for confirming the TXT is verified and the apex has no certificate yet. The 06:14 UTC check was after the rollback. During the earlier 02:47–02:57 UTC window, authoritative DNS and two public resolvers returned only the Railway edge IP for the apex. At 02:48 UTC, an HTTP request to that edge with Host haidaratelier.com got a 301 to https://haidaratelier.com/ with a Railway response header. Strict HTTPS still failed at 02:56 UTC. The old route is now back by design.

Could someone check whether Railway’s HTTP routing validation ran or failed during that window? Is there an official way to prevalidate this apex while the current site stays online, such as DNS validation or forwarding just the certificate challenge path? If the apex must move first, how can I trigger or observe validation on this binding when there is no retry button, and what status should I see before leaving traffic there?


Status changed to Awaiting Railway Response Railway • 12 minutes ago


Railway
BOT

11 minutes ago

We can't prevalidate the apex while the current site stays online. The certificate check for this domain uses an HTTP request to the domain itself, not DNS validation, so it only passes once the apex actually routes to nv32sez8.up.railway.app. Forwarding just a challenge path from the old host isn't a supported way to complete it, and a matching TXT on its own doesn't mean the domain is ready for cutover.

The most recent check we can see for this binding is the 06:14 UTC one, and it failed because requests weren't reaching us. That record doesn't tell us how any check during your 02:47-02:57 UTC window turned out.

To proceed, point the apex at the target and leave it there. Validation runs again on its own, and there's no manual trigger while the domain is still validating ownership. In the service's Networking settings, the apex is ready when it no longer shows as validating and has a certificate, and an HTTPS request with SNI for the apex returns a certificate for the apex instead of *.up.railway.app. We can't give a fixed time for that, so plan for HTTPS errors on the apex until it completes.


Status changed to Awaiting User Response Railway • 11 minutes ago


Welcome!

Sign in to your Railway account to join the conversation.

Loading...