Generally, if the performance issue is based on the System resources availability, then in a multi-node environment allow this particular agent to run in a dedicated background processing node that is classified [Node classification feature for a cluster environment].
Consider the scenario of a small application where there is only one production server. In this particular scenario, consider having multiple Standard Agents as part of the advanced agent processing. Kind of _Vertical Scaling of Agents in a single application server.
PEGA OOTB comes with SYSTEM-QUEUE-DEFAULT class. We can extend the SYSTEM QUEUE to have a custom queue class as well. The standard agent process the records present in this queue database table. QUEUE-FOR-AGENT method is used for the queue entry as part of the case processing.
