I'm fairly new to Weblogic 10.3 and java in general for that matter. Soon we'll be building a J2EE process utilizing JMS queues to process messages from another application. Its essential that our process handles, at most, 500 messages in one "batch", however we cannot process the next batch until this process finishes entirely. What we're doing is basically reading from a table, running some logic then sending messages to another application utilizing their webservices. We want to avoid rereading from that table and sending another batch of 500 to the message queue until the previous process is complete. Are there some standard practices I can look into for this.. hopefully this makes sense.

You can implement using service activator pattern.  Let the logic to generate message be the client to service activator and send messages to  external webservice.