Limiting Clusters/VMs to be monitored?

[FCOETech]

In our organization, we have a single DataCenter, and several clusters: one for Production servers, one for Development servers, one for VDI, one or two other small ones. I was recently asked to look into limiting the monitoring from vCOps, as our development cluster was constantly sending "noise" due to the way the machines are used by the programmers in that cluster. Management decided that they would prefer only the production cluster is monitored.

I did not perform the initial install of vCOps and am learning about the product as I go. I did review the install documentation and KB2078313 ( http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=207831... ) and also a few informal blogs outlying the same process as the KB. It sounded like the intent of this was more for preventing overuse of licensing when limited to a particular number of machines. However, it seemed it may work to limit monitoring as well.

I performed the changes in permissions on vCenter as noted, and then set up a collector account with this custom role, then only gave it permissions on the datacenter and single cluster we want to monitor. I changed the collector account in the admin of vCOps and tested connection.

I waited a good day to be sure the front end updated, but I am still seeing our entire datacenter on the dashboard of the management interface for vCOps.

Looking for some advice from anyone else who has attempted to limit monitoring in this way. Is there a better practice? Did you experience any hurdles not outlined in the KB? Thanks in advance!

[rcporto · Accepted solution]

Which version you're using ? 5.8.x or 6.x ? Anyway, instead of limit which cluster to monitor, why not create a different policy to your cluster and change the alerts settings and thresholds/limits ? This way you will continue monitoring all your environment without generate "noise"

[FCOETech]

Thank you Porto. We're on 5.8

I agree, this would be better. At the time it was outside of my scope of knowledge and I was in a bit of a time crunch initially to make this happen. Was hoping for an applied solution that would work without requiring me to learn a new system that I would not also be expected to manage. I have since had more time approved and been working within design to adjust the policies and create some exclusion groups. Still a bit tricky since the nature of most stuff in this cluster for us is to be broken by dev; vCOps doesn't really know if the state is down because the asset is being used for testing, or if it's down because a managing service took a dive. Or, maybe it does and I just haven't gotten to that level of management yet in the application! Either way, exclusion groups and lower thresholds are getting us much closer. I think what I was looking for and didn't ask clearly enough was more clarity on how to create and manage an exclusion policy for some datacenters/clusters/hosts/machines.

As a side note, the collector info did eventually update on the dashboard, so the posted solutions are applicable. Came in today and it was reflecting the correct information based on only what the collector account has access to. It just took an extra day to do so (and I'm not entirely sure what triggered it - I probably could have restarted services from the admin console to force this sooner?)

Thanks again for the feedback!