The reaction of this triger is to quick. I try to catch errors of GPU by reading a core clock property, if equals 0 then i do rig reboot. But after starting new mining process not all cards is initialized and triger will be activaed. This offten happens with NBMiner. Maybe some delay can be implemented in this triger? Adjusting Options->Advance->Ignore rules on startups not help for this triger.
Regards.
Can you please try to combine it with the trigger "Mining uptime"? This trigger can be used to ensure the mining software has been running for a few minutes. You can have these two triggers (Device Property + Mining uptime) and "Match all" configured in your rule.