How it works
How the monitoring works.
A short, honest walkthrough — the same material our administrators read.
Probe
Prometheus and blackbox_exporter do the checking; the app never forks a ping. Each address is probed on its own schedule, chosen per device, per map, or globally: interval, timeout and how many failures mean down. A profile change regenerates the scrape jobs and alert rules, and Prometheus reloads only when the generated content changes.
Confirm
Display status is a rollup of the latest probes. Alerting status is deliberately different: an address is down only when its rule is firing, up when a probe succeeds, and anything in between — pending, not yet evaluated, or no data at all — keeps its last confirmed state. That hold rule is what makes a profile change or a Prometheus restart a non-event.
Suppression then explains what is left. A device behind a down parent is unreachable and silent; a device behind a down tunnel waits for the tunnel's verdict; maintenance windows hold their own alerts and send one DOWN when they end, if the device is still down.
Notify
The app sends every message itself, from its own confirmed transitions — never from Prometheus changing its mind. Targets are assigned per map and per device; quiet hours hold per target; escalation adds people when nobody acknowledges; phone calls ring in order with keypress acknowledgement. Everything is logged, with the reason when nothing was sent.
Store
Configuration, alert state and events live in SQLite; probe history lives in Prometheus. Both run in the same Docker stack on one VM. There is no cloud dependency, no LibreNMS and no outside database.
What it is not
It is not a NetFlow or DPI tool, and it does not replace a ticketing system. It monitors reachability, SNMP health and the RouterOS operations an ISP actually runs — and it is designed to stand alone doing it.
See it for yourself
The live demo is a real install.
Made-up network, real probing, guest read-only sign-in. No sales form in the way.