I'm seeing a persistent outage on two custom domains under my team (nh-lacrosse) — nhlacrosse.com (project nhlacrosse) and nvelope.app (project nvelope) — while a third domain on the same account (nathanielhunt.com) works fine.
Symptoms:
HTTPS requests to the affected domains fail with a TLS connection reset during the ClientHello (before any certificate exchange). Plain HTTP (port 80) returns a 503 from the edge. The underlying deployments are completely healthy — hitting the *.vercel.app URLs for these same projects returns 200 with no issues.
What I've done:
Found and removed a leftover wildcard CNAME record at my registrar (*.nhlacrosse.com → uixie.porkbun.com) that was intercepting _acme-challenge.nhlacrosse.com. Confirmed removed by querying the authoritative nameserver directly. Removed and re-added the domain from the Vercel project, then re-created the alias to the current production deployment — twice, once before and once after the DNS fix. Even after the DNS fix, connecting directly to the edge IP (216.198.79.1) with the correct SNI hostname via curl --resolve still resets the connection. The SSL certificate for the domain shows the same cert ID and age from before all of this — it doesn't appear to have been reissued despite the remove/re-add cycles. No incidents reported on vercel-status.com.
This looks like a stuck edge routing/certificate state for these specific domains rather than a DNS or config issue on my end at this point. I'm on the Hobby plan, so I don't have access to a direct support ticket for this problem area — hoping someone from the team or community has run into this before. Happy to provide more diagnostic output if helpful.