Before the cluster was completely reconfigured the capacity was a certain amount (111.71 TB in the example below)
After the cluster reconfiguration the capacity shows greatly a reduced amount (27.95 TB in the example below)
vSAN OSA (all versions)
The cluster is network partitioned on the vSAN network due to one or more hosts having "esxcfg-advcfg -s 1 /VSAN/IgnoreClusterMemberListupdates" set.
This setting prevents vCenter from sending a unicast agent list update to the hosts. This list instructs vSAN which host IP addresses are cluster members. Without this list present, or with an incorrect list present, the cluster cannot form with all members.
This can be verified in the vSAN Skyline Health check showing an alarm for "Cluster Partition".
When the cluster is not formed with all hosts, the capacity calculations can only be gathered from members of a single partition. In the example given the cluster has 4 hosts, all of which are in their own partition. This resulted in the displayed capacity being approximately 1/4 of the expected amount (29.75 vs the expected 111.71).
On all hosts, disable the setting to ignore vCenter pushing cluster member updates with the command "esxcfg-advcfg -s 0 /VSAN/IgnoreClusterMemberListupdates"
Verify the setting with the command "esxcfg-advcfg -g /VSAN/IgnoreClusterMemberListupdates"
Alternative you can ensure all hosts are set to ignore the member list from vCenter and manually configure.