Right now when an environment has redundant CI's e.g. switches and you want to see everything upstream from the switch that will be impacted (e.g. level 2) you end up cycling out or if you wait long enough you get the same CI back in the results multiple times.
It’s not easy to show because the cyclic relationships break the view fuses too. One has to use a subgraph definition or self-relationships to get any possible number of layers. Redundant switches then loop back to themselves every third hop or so. The problem multiplies everytime one switch group connects to another. Think redundant Cisco Catalyst Switch Stacks. Here’s a picture that shows the problem in an over-simplified manner.
Need a way to only see the CI once, whether that is to create a cached list of CI's that it checks against when doing the anlysis and if the CI is in that list it does not show it again... I'm sure there are dozen ways to solve this, but it is a pretty common issue with our customers.