I've done something like this before with scripting. At the time we only wanted an alarm if both CPU utilization and processor queue length were over their thresholds. We had a script that would generate a new alarm if it found both a CPU utilization alarm and a processor queue length alarm for the same server. The original two alarms would not open tickets, but the new combined alarm would. That prevented us from reacting to the individual alarms.
Even though you would probably prefer a built-in or configurable solution, depending on how important it is to you, you can definitely pull it off with some time spent scripting.