I think systems like nagios are enough good for general monitoring.
Glusterfs itself is a part of storage, so a monitoring system should
monitor e.g. disk health/RAID status too.
Advanced monitoring
(requiring knowledge of glusterfs internals) could be useful e.g. to
define bottlenecks. AFAIK glusterfs developers welcome contributions so
somebody could be a pioneer by the question.