Prism Element and NCC are two ways to run all NCC checks manually on a Nutanix cluster. Prism Element is the web console that provides management and monitoring capabilities for a single Nutanix cluster. Prism Element has a Health dashboard that shows the status of various components and services in the cluster, such as disks, nodes, CVMs, NCC, and alerts. The Health dashboard also allows the administrator to run NCC checks manually by using the Actions drop-down menu on the right side of the screen. The administrator can choose to run all NCC checks or specific checks based on the category or severity. The NCC checks will run in the background and generate a report that can be viewed or downloaded from the Summary tab. This method is easier and faster than running NCC from the command line on the CVM.
NCC stands for Nutanix Cluster Check, which is a framework of scripts that performs system checks and validations on Nutanix clusters. NCC can detect issues related to hardware, software, configuration, hypervisor, networking, and more. NCC can be run from the command line interface (CLI) of any CVM in the cluster by using the ncc command. To run all NCC checks manually, the administrator can use the command ncc health_checks run_all, which will execute all available checks and display the results on the screen. This method is more comprehensive and detailed than running NCC from Prism Element.

A VM-VM anti-Affinity policy is a rule that ensures that two or more VMs don't run on the same AHV host.It's useful when an application provides HA and an AHV host can't be an application's single point of failure1. In this case, the SQL cluster VMs should have a VM-VM anti-Affinity policy configured to prevent them from running on the same host and causing an outage if that host fails.A VM-VM anti-Affinity policycan be created using the aCLI commands2. The other options are not relevant for this scenario.
References:1:Affinity Policies - Nutanix Support & Insights2:Affinity Policies Help | Nutanix Community

Zombie VMs and overprovisioned VMs are two types of inefficient VMs that can waste resources and increase costs in a Nutanix environment. Zombie VMs are VMs that are powered on but have no activity or utilization for a long period of time. Overprovisioned VMs are VMs that have more resources allocated than they actually need or use. Both types of VMs can be identified by adding predefined views to a report in Prism Central.
A predefined view is a template that defines what data is displayed and how that data is represented in a report. Prism Central provides several predefined views for different purposes, such as capacity planning, performance analysis, anomaly detection, and efficiency optimization. To add a predefined view to a report, go to Operations > Reports > New Report and select the desired view from the list1.
The Zombie VMs List view shows the list of zombie VMs in the environment based on the CPU usage, memory usage, disk IOPS, and network throughput metrics. The view also shows the amount of resources wasted by these VMs and the potential savings that can be achieved by deleting or resizing them2.
The Overprovisioned VMs List view shows the list of overprovisioned VMs in the environment based on the CPU usage, memory usage, disk IOPS, and network throughput metrics. The view also shows the amount of resources wasted by these VMs and the potential savings that can be achieved by resizing them3.
By adding these two views to a report, an administrator can identify inefficient VMs and take appropriate actions to optimize resource utilization and reduce costs.