Dependencies impact how devices are polled and the events that can get triggered.
Up Dependencies (i.e. these devices are ‘Up’ Dependent on: this device) are configured to prevent excessive notifications behind a single down device. For example, if the poller has to go through Router 1 to reach Routers 2-50 then, routers 2-50 (i.e. These Devices) should be “Up dependent on” Router 1. The result, if Router 1 stops responding to the poller, the administrator will be notified for Router 1 and not Routers 2-50. With Up Dependencies applied, it would be assumed by the poller that routers 2-50 are unreachable and therefore, polling to routers 2-50 would stop until Router 1 responds to a poll.
- Up Dependencies will stop the polling on all of the devices that are Up dependent on a down device. This logic helps maintain optimal availability reports on the devices behind the down device.
Down Dependencies (i.e. these devices are ‘Down’ Dependent on: this device) are configured to avoid polling unless a specified device stops responding to a poll. For example, Router B should not be polled unless Router A stops responding to a poll. This configuration would mean that Router B is ‘Down’ Dependent on Router A. This scenario is desirable in a high availability network configuration.
- Down Dependencies will start the polling on a device that is Down dependent on a down device. For example, if Router A goes down, Polling on Router B would commence until Router A responds to a poll.
Configuring dependencies is a best practice for scaling the network polling process while managing the volume of notifications sent when an outage occurs. If possible, they should be configured all the way back to the poller. This practice limits the risk of excessive notifications and aids the administrator in isolating the exact location of an incident.