Node local DNS cache: write: operation not permitted

[chconr]

We are trying to deploy node-local-dns to our EKS clusters.

Kubernetes version: 1.31
CoreDNS helm deployment: coredns-1.39.1
node-local-dns config overrides:

config:
bindIp: true (we are running bottlerocket as a base image)
commProtocol: "prefer_udp"
dnsServer: "172.20.0.10"

We are seeing the following failures in the node-local-dns daemonset pods:

[ERROR] plugin/errors: 2 github.com.cluster.local. A: write udp ->: write: operation not permitted
[ERROR] plugin/errors: 2 github.com.cluster.local. AAAA: write udp ->: write: operation not permitted
[ERROR] plugin/errors: 2 github.com.cattle-fleet-system.svc.cluster.local. A: write udp ->: write: operation not permitted
[ERROR] plugin/errors: 2 github.com.cattle-fleet-system.svc.cluster.local. AAAA: write udp ->: write: operation not permitted
[ERROR] plugin/errors: 2 linkerd-dst-headless.linkerd.svc.cluster.local. AAAA: write udp ->: write: operation not permitted

I am suspecting there are issues with permissions and communication to the CoreDNS deployment. I see references in the documentation to k8s:kube-dns and kube-system namespace requirements. Our coredns deployment is using k8s:coredns and coredns namespace. What's the best to get this working?

Thanks.