This paper describes the implementation of a monitoring system aimed for universities, enterprises or on prem hosting services. We survey the most relevant time series database systems and visualization tools available and selected a Prometheus time series database to collect the metrics from the equipment and to serve them to Grafana, a visualization tool that allows to build dashboards for a holistic observability. Then we integrated Prophet, a machine learning tool able to detect pattern anomalies in the collected data and trigger warnings automatically. We evaluated the system for performance and scalability and concluded it is very lightweight in resource consumption which means it could scale to large networks.