Contents

Overview of CPU load detection

You can monitor the CPU load on your CIC server in order to determine if IC Server’s Notifier and Queue Manager subsystems are overloaded.

To configure CPU load detection, you set a high watermark value, a low watermark, and the sample period. IC Server calculates the time it takes to send two back-to-back ping messages to Queue Manager during a sample period. It then divides this resulting number by two to get the average time. The average time indicates how quickly Notifier is passing messages, and if Queue Manager is keeping up with queue transitions or attribute changes.

IC Server then compares the average time against the high watermark and the low watermark values:

If the average time of the ping messages take longer than the high watermark, then IC Server sends a message to the CIC subsystems. This message states that the CIC system is experiencing high load conditions. When this occurs, e-mails are no longer queued, Interaction Dialer stops placing campaign calls, and other subsystems take action to reduce the load on the system.
IC Server continues to send ping messages. When the ping messages result in an average time that is less than the low watermark, IC Server sends a message to the CIC subsystems. This message states that the CIC system has returned to normal. The CIC subsystems then return to normal as well.

Note: In multiprocessor systems, the average ping time is measured across all CPUs. For example, if you have two CPUs and one is 50% busy and the other is 100% busy, the CPU utilization is 75%.