Stress Test on JMS Modular Input TA and UF

We are having a stress test done on JMS Modular Input TA and UF. Could you pl give us pointers to having the number of instances increased on both JMS TA and UF to scale it up in case the need arises. Topology we have: JMS Mod Input receives log messages from JMS Topic (native MQ Pub/Sub). JMS Mod Input passes them on the UF. On all the same server (MQ, JMS TA and UF). UF then pushes these logs to Splunk on another remote server. Indexes are configured on Splunk to view these logs and query on them