It’s possible they started using Azure Load Balancer internally as part of an architectural change. This might be for scaling ingestion endpoints, handling control plane traffic, or improving internal traffic routing. Unfortunately, there doesn’t seem to be any official documentation or announcements about this.
If you’re trying to dig deeper, a few things that might help:
Activity Logs: Check the Azure Activity Logs in the resource group to see if any backend resources are being provisioned or associated with the cluster.
Network Watcher: If enabled, use it to inspect traffic flows and confirm what’s going through the Load Balancer.
If it’s driving up costs, and you're not explicitly using a load balancer in your design, it’s definitely worth raising with support to see if there's an optimization or config workaround.
Would be great to hear what you find out.