We would like to reliably detect a situation where a Windows guest crashed with BSOD. In such cases, we would like to receive an alert from Prometheus. There are two ideas for achieving this:

1. The first method involves detecting when the QEMU guest agent fails to report. An example of a readiness probe that utilizes this approach can be found in the comments below. Note that this detection method is not entirely reliable, as there may be circumstances in which the QEMU agent does not respond for reasons other than a Windows BSOD.
2. Utilize the detection through the QEMU pvpanic device. The pvpanic device is a virtual device specifically designed to detect guest panic events. This would offer a more reliable solution.