A Custom Alert Policy should be created for the target VM’s CPU-utilization metric. The policy can define the VM entity, the CPU threshold of 90%, the duration or evaluation condition required to establish sustained utilization, and the alert severity. Prism Central evaluates enabled user-defined alert policies periodically and generates an alert when the configured conditions are satisfied.
Email notification additionally requires alert-email configuration. The administrator must define valid SMTP settings or the applicable notification channel and ensure the intended recipients receive alerts matching the policy. This separates metric evaluation from notification delivery: the Custom Alert Policy detects the condition, while alert-email settings determine how the resulting alert is distributed.
A storage-container quota monitors storage capacity rather than VM CPU utilization. Pulse transfers diagnostic telemetry to Nutanix Support and does not provide customer-defined VM CPU threshold policies. A Recovery Plan orchestrates disaster-recovery operations and is unrelated to performance-alert evaluation.
Using a sustained-duration condition is important because a brief CPU spike does not necessarily indicate a capacity or performance problem. The policy should reflect how long utilization must exceed 90% before notification is justified.
NHCF/NCA reference: Custom alert policies, metric thresholds, alert evaluation, and email notifications. Official references: User-Defined Alert Policies , Alert Metrics , and Configuring Alert Emails .
===============
