Identifying long-running and long-pending requests

A group of parameters help to identify requests that are running too long, or requests that are pending too long. In case a concurrent request is running longer than the threshold you have set, theNumLongRunningReq parameter will enter an alarm state. In case a concurrent request is pending longer than the threshold you have set, the NumLongPendingReq parameter will enter an alarm state.

Two parameters in the CONC_REQUEST (OA_REQUEST) application enable you to identify long-running and long-pending requests on all concurrent managers:

Parameter	Type	Description
NumLongRunningReq	Graph	Displays the total number of long-running requests that exceed the long-running threshold. You can review the information specific to a point on the graph by double-clicking the annotated data point. The annotated data point will reveal the long-running request name, the concurrent manager name, and when the concurrent request was started.
NumLongPendingReq	Graph	Displays the total number of long-pending requests that exceed the long-pending threshold. You can review the information specific to a point on the graph by double-clicking the annotated data point. The annotated data point will reveal the long-running request name, the concurrent manager name, and when the concurrent request was started.

Note

Although a long-running request might not be running long enough to cross the long-running threshold, it can run long enough for requests that are waiting for the same concurrent manager to cross the long-pending threshold. If a long-running request causes pending requests to wait long enough to cross the long-pending threshold, you can alleviate the issue by adding a dedicated concurrent manager or scheduling the request to run when the concurrent request activity level is low.

Identifying long-running and long-pending requests

Related topics

On this page