Use wsrep_sync_wait to detect replication lag in Galera clusters

The Galera monitor can detect the lag between certification and application of events in a Galera cluster by writing an event on the write-master node and doing "dirty reads" on the nodes labeled as slaves. This gives similar functionality for Galera clusters that is already in place for Master-Slave clusters in mysqlmon.

Another option is to use wsrep_local_recv_queue as a measurement of lag in a Galera cluster.