Master:
SELECT pg_current_xlog_location();
Slave:
SELECT pg_last_xlog_receive_location();
and by comparing these two values you could see if the servers are in sync.
The problem yet again is that if streaming replication fails, both of these
functions will continue to return same values and you could still end up
thinking the replication is working. But also you need to query both the
master and slave to be able to monitor this, which is not that easy on
monitoring systems, and you still don’t have the information about the
actual lag in seconds, so you would still need to run the first query.
postgres 
