A specific datastore reports 100% disk usage or triggers a "Datastore usage on disk" alarm.
The underlying storage array (SAN/NAS) also reports 100% allocation for the associated LUN/Volume.
The datastore contains a Virtual Machine (VM) often used for high I/O or backup operations.
Despite the Guest OS within the VM showing free space, the vSphere Client reports the datastore is fully consumed.
VMware vSphere ESXi 8.x
This issue is caused by the disk provisioning format selected during the creation of the Virtual Machine's hard disks.
The virtual disks attached to the VM are provisioned as Thick Provision Lazy Zeroed. unlike Thin Provisioning, Thick Provisioning reserves the entire allocated size of the virtual disk on the physical datastore at the moment of creation, regardless of how much data is actually written by the Guest OS.
Because the space is strictly reserved to guarantee performance and availability, vCenter and the underlying storage array correctly report this space as "consumed" or "used," resulting in the 100% utilization metric.
This is the expected behavior for Thick Provisioned disks (both Lazy Zeroed and Eager Zeroed). No immediate technical fix is required if the intent was to reserve all storage capacity for this VM.
To reclaim unused space: Converting the disk format to Thin Provision allows for the reclamation of unused space if the 100% utilization reporting is not desired.