Add disk name to disk_percent_used metric

[eduanb]

Currently, the only way to identify a disk in the disk_percent_used metric is by using the device tag, which is the low-level NVME device name like /dev/nvme1n1 . This isn't very user-friendly.

This device name can easily be mapped to the disk name in GCP by using the /dev/disk/by-id links. For example:

readlink -f /dev/disk/by-id/google-logs-disk
/dev/nvme1n1

[github-actions]

This issue was marked stale due to lack of activity. It will be closed in 14 days.

[eduanb]

This is not stale and is still a valid feature request.

[emanueleaina]

The device name is also largely meaningless in alerts: since it changes on every reboot, adding it to the notification does not tell you which filesystem is actually getting full.

It also breaks snoozing if the alerting VM is rebooted: the wrong disk will get unexpectedly snoozed which can cause important alerts to be missed a new alert will be triggered for the new name of the already reported disk.