I tried to find any similar question in the community discussion but failed.
We are currently evaluating TeamCity as CI platform to migrate our Kernel Mode Driver projects for Mac/Linux/Windows to. As our product is Kernel Mode driver, it could and actually does kill the test machines (BSOD, Kernel panic, etc.).
So I wonder if it is possible to develop a plugin that fits TeamCity's architecture and can try to reenable 'dead' agents by using ESX API to return test machines to snapshots and power them on using ESXi API after agents became non-responsive?