A Nutanix AHV VM was cloned from a template after a guest OS upgrade. Prism Central shows the VM's Nutanix Guest Tools (NGT) communication link as inactive, application-consistent snapshots of the VM fall back to crash-consistent, and in-guest self-service file restore does not work. Which condition should you validate first?

Nutanix Guest Tools is the in-guest agent and service that connects a VM to the cluster for features that need guest cooperation. Application-consistent snapshots rely on NGT to drive VSS on Windows or quiesce scripts on Linux, and self-service restore relies on NGT to mount snapshot disks inside the guest. When those features fail together and Prism shows the NGT link as inactive, the agent is the common factor. A clone gets a new VM UUID, so NGT must be enabled and mounted for the clone rather than inherited from the template, and an OS upgrade can stop or outdate the NGT services. A failed customization script runs once during first boot and would not break snapshots or file restore later. A SATA disk controller affects disk presentation and performance, not the NGT communication channel. A Flow policy that blocks ICMP stops ping, but NGT talks to the cluster virtual IP over its own TCP port, so ICMP has nothing to do with these symptoms. Note that AHV reports a guest's IP address through its network layer, not only through NGT, so a missing IP alone would not point to NGT. Exam caveat: map the symptoms to the feature that depends on NGT: application-consistent snapshots, self-service restore and in-guest scripts. Operational check: in Prism Central, check NGT status for the VM, enable or mount NGT for the clone if needed, and confirm the guest service is running.