I’m planning to use HAProxy and have some concerns about horizontal scaling. Our current setup involves using the Data Plane API to dynamically modify HAProxy map files, which store rate-limiting information that needs to be changed periodically. This works perfectly when deployed on a single Kubernetes pod, with the Data Plane API running as a sidecar container.
However, I’m wondering what happens when the load increases, and I need to deploy multiple HAProxy pods. Would I need to run a separate instance of the Data Plane API for each pod, or can one instance manage all HAProxy pods?