Netforge.ai Docs
Open the app →
Docs/Guide/Monitoring & Uptime

Monitoring & Uptime

Watchtower uptime monitoring, BGP route monitors, and external status.

Relevant for:Enterprise / NetOpsAdmin / Owner

Watchtower#

Where to find itTools→Watchtower

Built-in uptime monitoring: HTTP(S), TCP, ping, DNS, page-change, heartbeat and BGP checks on your endpoints with response-time history, incident timelines, and alerting via e-mail, SMS/voice (Twilio), Telegram, and browser push. Status pages can be shared with a token for stakeholders.

Why
The first minutes of an outage are triage: is it the app, the network, or the ISP? Independent synthetic checks with per-check history answer that instantly — and alert routing by severity keeps 3 a.m. pages for the things that deserve them.

BGP monitors#

A BGP monitor watches how the internet routes a prefix, an ASN, an IP or a domain, using the same RIPE route-collector data as Looking Glass — no packets are sent to your network. It alerts when the origin AS changes, when a second AS starts announcing your prefix, when RPKI or IRR registration changes, or when a blackhole community appears. A withdrawn prefix is the one condition reported as down; everything else is a change.

The optional expected origin AS is worth setting. Plain change detection compares against whatever was true when the monitor was created, so it cannot flag a hijack that was already underway at that moment. Naming the AS you expect turns that into an alert on first observation instead.

Note
BGP monitors check at most once an hour. The underlying collectors reconcile on roughly an eight-hour cycle, so a shorter interval would spend a third party's API budget without returning anything new. A routing-data outage upstream is recorded in the check history but never opens an incident — it is not your network being down.
Why
An origin change or a ROA going invalid is completely silent from inside your own network: your routers keep announcing, your monitoring keeps passing, and traffic is already going somewhere else. This is the one check here that watches from the outside, and it is the only way to find out in minutes rather than when a customer calls.

External uptime#

The platform itself is watched from outside (UptimeRobot / Uptime Kuma) so “the monitor is down too” has an answer — never rely solely on a system to report its own death.