Using the inbound integration polling and throttle intervals

Two configuration parameters are available for coordinating when the inbound integration work begins: pollinginterval and throttleinterval.

The primary controller is the polling interval (pollinginterval). Additionally, you can use the throttle interval (throttleinterval) parameter to reduce the impact on the server load or external resources.

The inbound integration performs work in the following sequence:

1. At the beginning of the polling interval, the integration polls for new work and creates new work items but does not start processing the work items.
2. PolicyCenter begins working on the work items.
   • If the ordered parameter is true, the work is ordered. In this case, the work happens in the same thread.
   • If the ordered parameter is false, the work is unordered. In this case, the work happens in separate additional threads with no guarantee of strict ordering.
3. After the current work is complete, the system determines how much time has elapsed and compares that value with the two interval parameters. The behavior is slightly different for ordered and unordered work.
   • If the work is ordered, the server finishes the existing work. Next, the server checks the elapsed total time since last polling. If the elapsed time is greater than the sum of the pollinginterval and throttleinterval, the server polls for new work immediately. Otherwise, the server waits until the sum of the pollinginterval and throttleinterval has elapsed.
   • If the work is unordered, the same time check occurs. However, the time check happens immediately after new work is created but does not wait for the work to be done. The work proceeds in parallel threads. When the last work item in a batch completes, the session is closed.