In order to get alarmed just based on the thresholds, you will need to simplify the rule, If you don't want to be alerted for spikes.
Note: It's better if you create a copy of the original Rule, disable the original and modify the copy, in case you need to reverse any changes. The changes can be done like this:
1. By going to Administration │ Rules & Notifications │ Rules
Having selected Cartridge: Virtual-VMware
2. Do a search for "VMW Virtual Machine CPU Utilization", click on it and choose "View and Edit" and click on "Rule Editor", then under "Conditions, Alarms & Actions" do changes for both "Critical" and "Warning"
You can replace the original logic with the below:
if(#utilization from $scope.hostCPUs # > registry("VMW:virtualMachineCPU.critical"))
Note: Virtual-VMware-Intelliprofile has another rule for CPU Utilization but but does not have thresholds.
You can also reference KB on how Standard deviation work: