Changes in the size of vsanDatastore indicate a portion of the cluster (e.g. disk/Disk-Group/node) is temporarily becoming unavailable - either due to intentional activities (e.g. placing a host in Maintenance Mode) or issues with the disks/controller/network.
In this case, judging by the size changes, the numbers indicate that 1 node in a 10-node cluster is becoming unavailable (or 1 Disk-Group in a 5-node cluster with 2 DGs per node).
If this is the same cluster you referred to here vSAN Health Alarm 'Hosts with connectivity issues then I would advise starting with the points I referred to which you provide any further information on.