Hi,

There is an issue with SMART monitoring for Samsung-based SSD.

Zabbix Agent 2 gets the RAW value from Smartctl Wear_Leveling_Count but on the SSDs from this brand, the RAW value is not pertinent, the desired value is the normalized one.

I've seen there is a turnaround for Linux hosts via a host template, but it's of no use for Windows hosts.

Is it possible to make Samsung SSD Wear_Leveling_Count support accurate by reading the normalized value instead of the raw one ?

Thanks in advance

Steps to reproduce:

• On host: do a smartctl on a Samsung SSD. The Wear_Leveling_Count Normalized and Raw values are what we want to see.
  

• On server: do a smart.disk.get, the value passed to Zabbix for the Samsung SSD is the RAW value, not the normalized one.
  

• On frontend: the value displayed is actually the raw value: (NB: in order to be able to get csmi0,4 from smartctl, I had to downgrade Agent2 to v. 6.4.8 because of bug ZBX-23789)