Azure Monitor Managed service for Prometheus now supports Horizontal Pod Autoscaling (HPA) for the ama-metrics replica set pod by default.
With this update, the ama-metrics replica set pod which handles the scraping of Prometheus metrics with custom jobs can scale automatically based on the memory utilization. By default, the HPA is configured to support a minimum of 2 replicas (which is the global default) and a maximum of 12 replicas. Customers will also the have ability to set the shards to any number of minimum and maximum replicas as long as they are within the range of 2 and 12. HPA will automatically take care of scaling based on the memory utilization of the ama-metrics pod to avoid OOM kills.