It looks like the issue may be related to resource utilization in the Istio Ingress Gateway pod. Check CPU and memory usage when the problem occurs, and increase the resource limits if necessary.

To avoid a single point of failure, always maintain a minimum of two Ingress Gateway pods.