Child group with two or more parents shows inconsistent state

Overview

You may have an environment in which two redundant routers are set up as two independent parent nodes in a dependency for a single group of nodes.  When only one of the routers goes down, the child group still triggers and consequently a flood of alerts is sent as the nodes miss pings but still do not go unreachable.  This is due to the fact that one of the parent nodes is still up.

Environment

All Orion environments

Cause 

Two parent nodes for one child group or node creates a conflict in status where the child group does not go unreachable due to having at least one parent in an up state.

Resolution

1. Create a group for the parent nodes and make the group the singular parent of the child group.
2. Set the parent group Status Rollup to Show Worst.  That way, when either of the nodes go down, the child node status is changed to unreachable as the singular parent goes down.
3. For more information, see the Using Orion Groups and Dependencies for best practices.