We have identified the cause of the ZFS replication issue.
There are large amounts of snapshots on the system causing degraded performance which caused a back log of jobs that are queued up and not competing.
We are currently purging the snapshots and limiting the amount kept to avoid this from happening in the future.