Stale NS delegation cache on one resolver node (dns1) — domain migrated nameservers days ago, one node still resolving via old NS

My linked-IP config resolves crm.kilnsidefarm.co.uk and news.kilnsidefarm.co.uk as NXDOMAIN when queried via dns1.nextdns.io (45.90.28.201), but correctly via dns2.nextdns.io (45.90.30.201) — same config, both queried directly.

The domain kilnsidefarm.co.uk moved from AWS Route 53 to Cloudflare nameservers as part of a planned migration. The .co.uk registry and dns2 both correctly show the current delegation:

kilnsidefarm.co.uk  NS  frida.ns.cloudflare.com
kilnsidefarm.co.uk  NS  terin.ns.cloudflare.com

But dns1 is still returning the old delegation:

kilnsidefarm.co.uk  NS  ns-1325.awsdns-37.org
kilnsidefarm.co.uk  NS  ns-1621.awsdns-10.co.uk
kilnsidefarm.co.uk  NS  ns-249.awsdns-31.com
kilnsidefarm.co.uk  NS  ns-634.awsdns-15.net

crm/news only exist in the new Cloudflare zone (created after the cutover), so dns1 returning NXDOMAIN for them is consistent with it still walking the stale AWS delegation. The root domain resolves "correctly" via both paths only because the legacy AWS zone happens to still carry the same old website A record.

This has persisted for 24+ hours, well past the zone's SOA negative TTL (1800s) and the parent delegation TTL, so it looks like a stuck/stale cache on that specific node rather than normal TTL-driven staleness. Could someone purge the delegation cache for this domain on dns1?