My production custom domain is stuck on an obsolete Heroku IP even though the Vercel deployment and dashboard configuration are correct.

Domain: contracktor.app Working deployment: https://contracktor.vercel.app

What I verified:

• The registrar and nameservers are Vercel (ns1.vercel-dns.com and ns2.vercel-dns.com).
• Both authoritative nameservers return 18.204.152.241 twice with TTL 10 for the apex.
• www.contracktor.app also returns 18.204.152.241.
• HTTPS to both custom hostnames fails during the TLS handshake.
• contracktor.vercel.app loads normally and returns HTTP 200.
• The managed apex ALIAS shown in Vercel is d0e8b97178b92aa7.vercel-dns-017.com; that target resolves to 216.198.79.1 and 64.29.17.1.
• Forcing contracktor.app to either of those Vercel IPs returns HTTP 200 from the correct deployment.
• Detaching and reconnecting contracktor.app to Production did not change the authoritative answer.
• Explicit apex A overrides for 216.198.79.1 and 64.29.17.1 appear in the Vercel dashboard but still are not published after more than a day.
• Other newer records (_dmarc, send MX/TXT) publish correctly.

The working .vercel.app hostname confirms the deployment is healthy; the failure is isolated to Vercel’s authoritative apex/www DNS routing or ALIAS flattening.

Could a Vercel team member purge or regenerate the stale managed apex and www records without disturbing domain registration or email DNS records?