There are a wide variety of ways to do this.
If your naming scheme is consistent, you could include/exclude devices based on that (where node caption contains/does not contain "SQL"
Or if the devices you want to include/exclude were part of the same subnet, you could do the same kind of contain/not contain on IP_Address
Or you could set up a custom property for all servers, and then filter by that value.
Or you could create a custom property for CPU threshold, and only alert when the CPU value is greater than the specified threshold. Then set devices you want to ignore to an impossibly high threshold (300%).
And much much more!
Was hoping to not have to modify the existing server AD properties to get a grouping done in a third party application. Thanks though, I may have to setup a custom AD field.
You might try posting this question over at this forum: Network Performance Monitor
Alert Central is a separate free product from NPM Orion. See Alert Central - Free Alert Management Software by SolarWinds