I have a Kubernetes (K8s) cluster on azure (AKS). I configured an ingress controller (Ngnix) to make my service available publicly. AKS relies on Azure DNS Zone to create an "A" record that maps to a Public IP Address generated by Azure. The issue is that after a couple days the IP Address disappears and the Ingress stops working.

My work around is currently to delete the following pods each time it happens:

kubectl delete pod  addon-http-application-routing-external-dns-XXXX -n kube-system

kubectl delete pod  addon-http-application-routing-nginx-ingress-controller-xxxxxx-n kube-system

Does anyone know why the IP gets lost each time? Is there a permanent fix to that?

Thanks!