Methods and a computing device are disclosed. A computing device may include a managed device having embedded firmware. When a failure occurs with respect to the managed device, drivers within the computing device may collect failure data from a driver stack of the computing device and from the managed device. The computing device may send the collected failure data to one or more second computing devices to be stored and analyzed. The computing device may include a health monitor for periodically collecting telemetry data from the computing device and the managed device. When the health monitor becomes aware of conditions indicative of a possible impending failure, the health monitor may trigger collection of sickness telemetry data from the computing device and the managed device. Collected data from the managed device may be made available to a vendor of the managed device.