I set up monitoring across estates of mixed capability — full SNMP where the device supports it, reachability and response-time tracking where it doesn't — with per-device history, exportable charts, alerting that fires on trend rather than only on failure, and packet capture wired in.