For SAS/SCSI disks, smartctl -x provides useful health and transport information which is currently not exposed by smart.disk.get[].

Example current output:

{{zabbix_agent2 -t smart.disk.get[/dev/disk/by-id/scsi-...,scsi]

{ "critical_warning": 0, "disk_type": "hdd", "exit_status": 0, "media_errors": 0, "model_name": "SEAGATE ST6000NM0095", "percentage_used": 0, "power_on_time": 55326, "temperature": 34 }

}}

However, smartctl -x for the same SAS disk additionally provides useful values such as:

• Elements in grown defect list
• Uncorrected read errors
• Uncorrected write errors
• Non-medium error count
• Percentage used endurance indicator for SAS SSDs
• SAS PHY statistics:
  • Invalid DWORD count
  • Running disparity error count
  • Loss of DWORD synchronization count
  • PHY reset problem count
• SMART self-test result / last extended test / first failing LBA

It would be useful if these values were exposed by smart.disk.get[] and added as dependent items/item prototypes to the official template.

For cumulative counters such as non-medium errors and SAS PHY errors, triggers should preferably detect an increase/change, rather than simply checking whether the absolute value is greater than zero.

Example:

Non-medium errors: 115 -> 116

is useful information, while permanently alerting on 115 > 0 is not.

Expected result

Better SAS/SCSI monitoring for enterprise HDDs and SSDs, allowing Zabbix to detect:

• media degradation;
• uncorrected I/O errors;
• increasing defect counts;
• SAS SSD endurance;
• SAS link/backplane/expander problems;

before the overall SMART health status changes.

I have attached the full smartctl --json -x output from a SAS HDD and a SAS SSD, together with the corresponding smart.disk.get[] output, to make it easier to compare which information is available from smartmontools but currently not exposed by the Agent2 SMART plugin.