Hey everyone,
I started to look into dependenices, and it appears all of our auto generated dependencies tend to be along the switch/router/firewall side of things. Has anyone had any luck with dependencies and virtual machines? Ideally the parent object would be a cluster or an individual host (test environment, non-ha type setups) and the child objects would be the nodes of virtual machines.
If the parent was an individual host, that side seems straight forward. Add the host as the parent item. I think if it was a DRS/HA cluster, you would most likely create a dynamic group where the status rollup is set to best status.
One of the main items I can't seem to wrap my head around would be the child items. Adding the virtual machines manually would be a management nightmare for a maintaining perspective. I'm not sure if there is a good Powershell, API, or SWQL approach to maintain said list. I thought about trying to do a group and have the various objects linked that way, but I'm not sure how that would affect the alerts. My understanding would be the alerts would need to be targetted to the group itself, where ours are targetted to the nodes.
Also apologies in advance, I wasn't sure of the best place to put a question about dependencies. I will happily move it if there is a better place for the question.
So I guess the TLDR question is:
What approach do you take to make a dynamic list of child objects in a depdency?
Scenarios:
Single host in HA cluster goes down (other hosts in cluster are still up): VM nodes still go to down state if problem
Single Host (Non-HA environemnt) goes down: VM nodes go to unreachable state
All hosts in a HA Cluster goes down: VM nodes go to unreachable state (I.E. suppress all but the couple critical alert actions avoiding a notification storm)