How do you troubleshoot and optimize monitoring and trending systems like Prometheus and Grafana at scale?
