Steps to reproduce:

1. SMART monitoring of (Seagate) SSDs
2. smartd detects "Prefailure Attribute: 1 Raw_Read_Error_Rate changed from 76 to 77"
3. smartd starts scheduled Short Self-Test
4. Self-Test takes few seconds
5. zabbix agent collects data while Self-Test in progress

Result:
Trigger "SMART {#NAME}: Disk self-test is not passed" evals to true immediately
Expected:
Trigger "SMART {#NAME}: Disk self-test is not passed" evals to true if Self-Test fail persists

I've currently changed the trigger expression to:

count(/SMART by Zabbix agent 2 active/smart.disk.test{#NAME},#2,"eq","false")>1

Same for non-active.