Hi, I'm trying to figure out how to attach fewer labels to timeseries stored by prometheus as described in https://docs.ranchermanager.rancher.io/how-to-guides/advanced-user-guides/monitoring-alerting-guides/debug-high-memory-usage#reducing-memory-bursts. Is there a way to exclude labels when using the helm chart for the rancher monitoring stack (grafana/prometheus/alertmanager)?