As 'rebalance' and 'remove-brick' are data migration operations, triggered from inside the filesystem itself, the quota usage limits are not enforced properly during that time.
During the time of rebalance/remove-brick it appears that the quota usage limits could have been not enforced as there is a chance that limits can be crossed because of temporary file creations by these processes.
Workaround (if any):
None. Ignore quota while these process are running.
Peace of mind :-)