Managing Dependencies

Dependencies are used to minimize false alarms: It does not make sense to be notified for several failing sensors that bello network monitoring cannot access on a server if actually the connection lines to the server are failing - then it is simply not possible for bello to tell whether the services are down or not.

To ensure that no false alarms are sent, the user can select "dependencies”—notifications are not sent if a selected sensor is down.

For example: bello network monitoring monitors three sensors on a server that are connected to the internet via a DSL line. It monitors the server by PING every 60 seconds and by HTTP request every 5 minutes.

If you select the PING sensor with a shorter interval as the dependency sensor for the two HTTP sensors, then you will not be notified for the failing HTTP sensors if the server is not reachable by PINGs, because the PING sensor will be marked DOWN in a matter of 1-2 minutes upon a failure.

You could also monitor the last trace route hop just before the server using PINGs. If the last hop is not available then the server can’t be available either and sending extra notifications for the failing webserver is not necessary since you already have received the notitications for the failing PING sensor.

Note: Use the TRACERT command to find out which is the next hop after your leased line.

More:

Notification Dependencies