While the cornerstones of efficient monitoring and alerting have been the same for ages, they are rarely applied in a consequent and consistent manner. The first half of this talk will give a perspective on how to do monitoring with minimal human interaction to avoid overload and/or pager fatigue. The second half will focus on how much more powerful and adaptive Prometheus is compared to legacy monitoring solutions. To wrap up, we will look at how pretty and useful Grafana is in combination with Prometheus.