NextDNS NA anycast routing broken at Misaka-IAD and Anexia-NYC

Anycast routing appears broken at your NA PoPs reached through Misaka-IAD (Ashburn) and Anexia-NYC. 

45.90.28.0/24 and 45.90.30.0/24 are unreachable from AWS us-east-1 on both UDP/53 and TCP/443. EU vantages on the same configuration are unaffected. Started ~08:22 UTC 2026-04-29.

Same host, same minute — only NextDNS prefixes fail

TCP/443 via curl --connect-timeout 3 (3 trials each):
1.1.1.1 [301/0.020s] [301/0.018s] [301/0.024s]
8.8.8.8 [302/0.015s] [302/0.014s] [302/0.016s]
9.9.9.9 [404/0.008s] [404/0.008s] [404/0.009s]
45.90.28.0 [000/3.002s] [000/3.002s] [000/3.002s] ← timeout
45.90.30.0 [000/3.002s] [000/3.002s] [000/3.002s] ← timeout

UDP/53 dig: same picture — non-NextDNS resolvers answer in <50ms,
45.90.28.0/30.0 timeout 3/3.

Rules out: VPC egress block, host firewall, port filtering. Selective to NextDNS anycast prefixes.

Traceroute terminates at NextDNS NA upstream colos

NextDNS NA anycast routing broken at Misaka-IAD and Anexia-NYC
to 45.90.30.0 — last responding hop: 94.16.100.32 (ae0-0.dcr01.anx32.nyc.us.anexia-it.net)

Packets escape AWS, traverse public Internet, reach Misaka-IAD and Anexia-NYC, then drop. From EU vantages traceroute completes at the anycast IP itself in ~6ms via European peering.

Ask

1. Are the `45.90.28.0/24` and `45.90.30.0/24` anycast announcements healthy on the NA PoPs reached via Misaka-IAD and Anexia-NYC?

2. If a PoP is degraded, can the announcement be withdrawn or fixed so transit best-paths to a healthy PoP?

3. ETA on remediation, if known.