Connect-and-handshake reachability for the service behind a host:port.
Every TCP monitor runs on the same spine as the rest of PulseTrace — probed from each region you pick, scored on every check, and streamed into one timeline where incidents open on their own, anomalies surface without a threshold to guess, and reports are generated for you. Point it at a target, choose an interval, and it's watched.
What each check verifies
The panel on the right is the real create-monitor screen — it fills itself in with a TCP example, exactly what you'd type.
For services that speak something other than HTTP — databases, message brokers, mail servers, SSH, custom daemons. It confirms the port accepts connections from each region and records connect time, without needing to understand the protocol.
postgres-primary:5432 is DOWN — connection refused from all 3 regions.