I don't this can already be fulfilled by combining relative and absolute conditions within an alarm rule.
I have created an alarm definition like you described (specifically VM State == Running AND
Guest Disk Free Space Relative < 10% AND
Guest Disk Free Space Absolute < 100GB).
One would expect that the alarm would only be triggered for disks that meet all three conditions, i.e. the disk:
- Is attached to a running VM
- Has less that 10% free space
- Has less than 100GB free space
However this doesn't seem to be the case. Consider the following scenario:
- VM is Running
- VM has multiple disks
- Disk 1 (C:\) has 50.50GB/99.90GB free (52%), which meets criteria 1 and 3, but not 2
- Disk 2 (D:\) has 134.44GB/1497.87GB free (9%), which meets criteria 1 and 2, but not 3
Because neither of these disks satisfy all three conditions (remember, the alarm is defined with AND
, not OR
) I would not expect the alarm to trigger, however the alarm does
trigger and Veeam ONE reports:
Amount of free disk space "D:\" (9.0%) is below a defined threshold (10.0%). Current free disk space is 134.44 GB.
Amount of free disk space "C:\" (50.5 GB) is below a defined threshold (100.0 GB).
State of virtual machine equals "Running"
I raised support case 02455134 about this, and the representative told me "I have contacted our seniors on this and basically, it is currently by the product design, so the final object is a VM and not the VM's disks, so if you have both conditions true for the VM - the alarm will be triggered, despite the conditions are true for different disks."
The representative also suggested that I post in the forums, and request that the check is changed so that the alarm conditions are evaluated per-disk instead of per-VM.