You can usually start multiple agent servers as well as increase
the threads in an agent JVM for increased throughput as long as there
are tasks in the queue for the threads. The agent finds work using
its getter, as described in Agent getters.
For each work, the agent then puts the work into the message queue
as tasks. As a result, the agent’s throughput is proportional to how
fast the agent can find and put tasks into the queue.
Sterling Selling and
Fulfillment Foundation logs
the following metrics in the YFS_STATISTICS_DETAIL table for each
agent:
- GetJobsTime - the total amount of time spent getting work
- MessageBufferPutTime - the total amount of time putting tasks
into the message queue
- ExecuteMessageCreated - the total number of tasks put in the message
queue
- GetJobsProcessed - the number of times the getter is executed
- ExecuteMessageSuccess - the total number of tasks that are successfully
executed.