Each agent is capable of processing five tasks at one time. Increasing the number of agents allows more tasks to be processed at one time, increasing the number of Data Mover jobs processed at one time. In some cases, this improves individual Data Mover job performance.
Data Mover jobs that use Teradata PT API or Teradata JDBC have separate tasks for each table being copied. Increasing the number of agents improves the job performance by allowing more of the job tables to be copied concurrently.
You may also want to limit the maximum number of current tasks that an agent can run by setting the agent.maxConcurrentTasks parameter in the agent.properties file. This parameter is set dynamically and does not require that the agent service be restarted. If you update the value, it takes effect one minute after the updated file is saved. In addition, you can also use flexible scheduling to set different concurrent limits depending on the time of the day.