We currently use dependencies to reduce unnecessary alerts. However, we have a number of devices that we need to count as down in our availability reports regardless of if the parent device is down.
Most displays and reports can be modified to include other statuses besides down (status=2). You can include unmanaged (status=9) and Unreachable (status=12). So I'm not sure this is needed as an enhancement.
I'm talking about the availability reports that calculates the percentage of time that the node is down. The information is stored in the ResponseTime table in the orion database. When a node is down, the majority of fields a entered as null, but the availability field is filled in as 0 for each polling interval,
When you have a dependency and the parent node is down, the entries in the ResponseTime table are omitted. This is where the built in availability report pulls its data. I'm sure that somebody who is good with SQL could potentially overcome this problem.
Here's an example of why we need this. I work for a financial institution and we have a group who responds to issues with ATMs. Using dependencies, we are able to send them an alert that informs them if the site is down (router down) or if the ATM hardware is down. In addition, we also track the availability of our ATM's for management. When the parent is down, due to the dependencies, the ATM is considered up. If I remove the dependency, I am no longer able to send my specialized alerts.
