fix: mark all devices sharing a physical GPU unhealthy on Xid errors

[tryuuu]

Fixes #1929

checkHealth kept only one *Device per physical GPU UUID in parentToDeviceMap, so devices sharing a physical GPU overwrote each other. As a result an Xid event marked at most one device as unhealthy. With replication (time slicing or MPS) the other replicas of the broken GPU stayed schedulable, and with MIG the event could even be dropped entirely because the targeted instance was never stored in the map.

This PR changes the map to map[string][]*Device and fans each Xid event out over all devices under the event UUID, so every affected device is reported as unhealthy.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[tariq1890]

Thank you @tryuuu for your contribution. On an initial glance, this looks fine to me. Let me run some more tests against this branch and get back to you sometime next week