An administrator wants to monitor virtual machines on a host and send notifications when memory usage reaches 80%. What should the administrator create in vCenter Server to accomplish this?

Host alarms are scoped to host-level metrics rather than virtual machine metrics, and the terminology is inverted - email notification is an action, not a trigger, in the vSphere alarm model.

A vCenter Server alarm is the correct object type for monitoring virtual machine metrics such as memory usage across a host; the 80% threshold is defined as a trigger condition, and sending email is configured as the action that fires in response - vCenter Server alarms support both VM-level metric monitoring and email action integration, making this the complete and correctly structured solution.

Host-level alarms are not the appropriate construct for monitoring individual virtual machine memory metrics; the correct object is a vCenter Server alarm, even though the action terminology used here is correct.

While a vCenter Server alarm is the correct object type, email notification is configured as an action (the response to a fired alarm), not a trigger (the threshold condition itself), making the terminology incorrect.