We use Tag-based Application Services only.
This works fine for automatically adding the infrastructure CIs belonging to the application, but other common CIs are naturally not added, i.e. nobody is aware if e.g. some network component fails.
I can solve this by grouping the relevant network CIs with a Dynamic CI Group, add the necessary key_values, and recalculate the AS. This way the AS goes red if something in the DCIG goes red.
But, the application might be running just fine on its servers. It might be the users that can not access the application, so maybe I should rather create the relationship from the application's Offering(s) to the DCIG's Offering(s). I believe the latter is more common to model manually (the tag-way is a bit of a "hackaround", but quite easy to do).
However, if we did top-down discovery, network components would be added to the map, and would thus make the AS go red, so I believe it makes sense to do an AS depends on DCIG.
Any opinions on this?
