Context:
My device monitor kept flagging ten devices it couldn’t name. They were my own Docker containers, but finding that out showed me the network had no memory. Netdisco now runs on its own VM and reads the router’s ARP table over SNMP every hour, so every address has a first-seen and last-seen record. A blackbox exporter pings all 254 addresses every minute for live up/down status, and Grafana joins the two into one board with a row per address zone. Along the way it disproved an address overlap I’d believed in for weeks.
This is the short version. The full story covers the ten unknowns, the overlap that wasn’t, and why last seen doesn’t mean alive.