Project: ruproa (Pro plan, Maericks Credit Builder team) Domain: ruproa.com --- apex on Vercel nameservers, dashboard shows Valid Configuration Cert: Let's Encrypt wildcard, issued May 26 2026, valid through Aug 24 2026
We're getting customer reports of SSL protocol errors loading ruproa.com. After investigation, I've found two separate issues and want to flag the Vercel-side one here.
VERCEL EDGE TLS FAILURES
SSL Labs external scan shows some edge IPs in the anycast pool fail the TLS handshake entirely while others serve the cert fine:
Working: 216.150.1.65 --- TLS OK, valid *.ruproa.com cert Working: 216.150.16.193 --- TLS OK, valid *.ruproa.com cert Failing: 216.150.1.129 --- "Failed to communicate with the secure server" Failing: 216.150.1.193 --- "Failed to communicate with the secure server"
Different public resolvers return different IPs from the pool (confirmed via Google 8.8.8.8, Cloudflare 1.1.1.1, Quad9 9.9.9.9), so customers land on different edges depending on geography/resolver. Those hitting the failing edges get ERR_SSL_PROTOCOL_ERROR.
Dashboard shows Valid Configuration for ruproa.com, www.ruproa.com, and ruproa.vercel.app. I clicked Refresh on both custom domains. The failing edges did not recover.
SEPARATE OPENDNS ISSUE (for context)
Additionally, Cisco OpenDNS/Umbrella has incorrectly classified ruproa.com as phishing and redirects to hit-phish.opendns.com. This is a false positive on Cisco's side (Talos rep is Neutral, Google Safe Browsing is clean, PhishTank has no listing). I'm pursuing that with Cisco separately. Mentioning it here in case Vercel has seen this pattern or has a channel to Cisco for domain reputation.
Has anyone else experienced partial edge TLS failures with Valid Configuration showing in the dashboard? Any guidance appreciated.