The performance issue is caused by Host CPU overcommitment, specifically regarding CPU scheduling (Ready Time) rather than total utilization.
While the Host CPU utilization is below 80% (meaning the processor has idle cycles), the configuration of 8 vCPUs per VM combined with a high 9:1 vCPU-to-pCPU ratio creates a bottleneck known as "Co-stop" or high "CPU Ready" time. The hypervisor scheduler must wait until 8 physical cores are available simultaneously to execute a single clock cycle for one VM. In a dense VDI environment, finding 8 free cores at the exact same moment is difficult, causing the VMs to pause (freeze) while waiting for the scheduler, even if the overall CPU usage of the host appears moderate. Reducing the vCPU count per VM would likely resolve the performance issues.

Comprehensive and Detailed Explanation From Exact Extract of Nutanix End User Computing documents:
According to Nutanix AHV Best Practices for Virtual Desktop Infrastructure (VDI):
* Power Management: The most critical "optimal" configuration for latency-sensitive EUC workloads is to ensure the host power management policy is set to High Performance. This prevents the CPU from entering power-saving states (C-states/P-states) which can introduce latency during user interaction.
* Memory Configuration: While Nutanix AHV typically disables memory overcommit by default (or does not use it in the same way as other hypervisors for VDI), the active configuration step required by an administrator to guarantee optimal performance is the Power Management setting. Memory reservation should generally be 100% (not 50%) to avoid swapping, and CPU/Memory shares are usually left at default unless there is specific contention, but Power Management is the baseline requirement for EUC optimization.
* Why not D? While memory overcommit should be disabled for VDI, in many AHV versions this is the default behavior or inherent architecture. "High performance power management" represents the specific tuning parameter that administrators must validate or configure (often in the BIOS or via AHV command) to ensure the environment is "optimized" for desktop workloads.

To ensure high availability for distributed applications such as load-balanced web servers, Nutanix AHV allows administrators to configure Anti-Affinity rules.
An Anti-Affinity rule specifies that a group of virtual machines should never run on the same AHV host simultaneously. By applying an Anti-Affinity rule to the three web server VMs, the Acropolis Scheduler ensures they are distributed across different physical nodes in the cluster.
In the scenario described, the web servers likely resided on the same host (or the failing host held the only active instance). If Anti-Affinity rules had been configured, the other web server VMs would have been running on healthy hosts, ensuring continued user access to the VDI environment despite the single node failure.

one of the possible causes of slower logon and application load times for VMware Horizon-based virtual desktop users is Failed to create Native clone errors. This error occurs when there is a problem with cloning virtual machines using Nutanix Native Clones technology. To troubleshoot this issue, you can check the following:
* The Nutanix cluster has enough free space to create clones
* The Nutanix cluster has enough CPU and memory resources to handle cloning operations
* The Nutanix cluster is running a compatible version of AOS and AHV with VMware Horizon
* The VMware Horizon environment is configured correctly to use Nutanix Native Clones
* The VMware Horizon agent is installed and updated on the master image
https://www.nutanix.com/support-services/training-certification/certifications/certification-details-nutanix- certified-professional-ncp-euc-v6
https://kb.vmware.com/s/article/1008360

In the terminology of Flow Network Security Next-Gen (FNS NG), the components are defined as follows:
* Security Policy Model: Defined as "A schema of policies for specifying and enforcing a desired behavior.4 A Policy Model will have one or more policies."5 This model provides the framework (such as Application, Isolation, or Quarantine) within which specific policies are created.
* Security Policy: Defined as a "collection of security rules and assets [entities, endpoints, categories, applications, etc.] on which the rules have to be enforced together."6 Therefore, the function that acts as the overarching schema and contains "one or more policies" is the Security Policy Model.7