Claymore, it is possible to add on remote monitor, a function to detect which gpu is hanging or restarting miner? (fault Gpu)
The miner has a watchdog, which will restart once a GPU hangs. In the remote monitor you can force restart the miner aswell, but the restarting on GPU hang is basically a default option.
You can set a -minspeed value to make the miner restart if the total MHASH is lower than the expected. (I run 354mhash and have minspeed of 350)