zpool reports a different total size than zfs

Taking SN05 as an example:

EVT22200005 # zfs list -o name,used,avail oxp_3b70cd03-9bf6-4d78-867b-5e58b3a1f3ff
NAME                                       USED  AVAIL
oxp_3b70cd03-9bf6-4d78-867b-5e58b3a1f3ff  2.24T   587G

EVT22200005 # zpool list -o name,size oxp_3b70cd03-9bf6-4d78-867b-5e58b3a1f3ff
NAME                                       SIZE
oxp_3b70cd03-9bf6-4d78-867b-5e58b3a1f3ff  2.91T

zfs list reports that the dataset's USED + AVAIL is about 2.81T, where zpool list reports 2.91T.

From zpool(1):

 The space usage properties report actual physical space available to the
 storage pool.  The physical space can be different from the total amount
 of space that any contained datasets can actually use.  The amount of
 space used in a raidz configuration depends on the characteristics of the
 data being written.  In addition, ZFS reserves some space for internal
 accounting that the zfs(8) command takes into account, but the zpool
 command does not.  For non-full pools of a reasonable size, these effects
 should be invisible.  For small pools, or pools that are close to being
 completely full, these discrepancies may become more noticeable.

The sled-agent uses the zpool list size as part of the zpool section of inventory, which written into the inv_zpool table, and eventually is consulted during region allocation (among other things).

Start with the sled-agent code that collects zpool data and writes the inv_zpool table, then compare the zfs and zpool capacity values shown in the issue. Trace how inv_zpool is consulted during region allocation. Before changing anything, establish which capacity should be used and verify the chosen value through the inventory and allocation paths.