About postgres scale out

Hello,

I'd like to ask a common question about scale out for postgres.

Our current data volume is about 500GB ~ 1TB in one pg cluster(postgres 9.2). We've set up master/slave replication to
keepsync.  
To reach better performance from apps side,  we want to use pgbounder like app to split read/write on multiple servers
insteadof single server. 
Sometimes, there is repl lag between master and slave pg servers (e.g. backup or other issues unexpected). Is there
flexiblesolution from pgbounder (or other ways) to identify that and switch read/write to not staled server smartly? 


Thank you in advance!