Your server is live. So what are you watching?
Every server in the network is watched around the clock. The panel shows you the same verdict we act on — reachability, change against the server's own history, and the controls to take it out of rotation without losing anything.
Reachability, in three windows
A server that answers its own heartbeat is not necessarily reachable from the internet — we learned that the hard way when a firewall rule silently cut a server off while it kept reporting itself as healthy. So reachability is measured from the outside: our API server probes yours, every minute, and the panel shows the result over 24 hours, 7 days and 30 days.
The probe checks whether your server answers, not how fast it is. It is not a speed test and it does not consume your quota.
Maintenance without losing your place
Pause takes the server out of the public network: it stops accepting new tests and finishes the ones in flight. Your own widget tests keep running. Resume whenever you're ready — the record, the history and the reports stay intact.
Decommissioning is the permanent version, and it drains first: five minutes of no new tests while running ones complete.
Reporting
Test volume, provinces, ISPs and the daily series live in Reports. Client IP addresses are masked; the panel shows aggregates, not individuals.
Active session counts and load figures are not in the partner panel — not on screen and not in the API response. That is operational data about the platform, not about the quality of your service.
Before you start
Network, hardware, DNS and reachability criteria are listed in server requirements.