z/OS® Message Flood Automation addresses the problems of message flooding on z/OS. z/OS Message Flood Automation does not claim to identify all cases of erroneous behavior, nor to take the 'correct' action in all cases. Its intention is to identify runaway WTO conditions that can cause severe disruptions to z/OS operation and to take installation-specified actions in these cases.

Message flooding causes disruptions as follows:

• Large numbers of messages to the z/OS consoles can obscure important messages and delay them from being acted on.
• Large numbers of messages to the automation system (for example, NetView®) can delay the processing of normal messages.
• Messages can use excessive CPU and storage resources. Buffering excessive message traffic can use large amounts of virtual and real storage and it can cause SQA to overflow into CSA. This can cause jobs, subsystems and complete systems to be delayed or even to fail.

Message Flood Automation can react to potential message flooding situations in a matter of tens or hundreds of messages (specifiable by the installation), before buffers begin to fill, console queues begin to build, and console message rates begin to skyrocket. Furthermore, its actions do not result in residual buffers or queues of messages that must be worked down to return to normal processing. Because its processing is targeted to the messages causing the problem, very few uninvolved messages are acted upon. By contrast, the act of flushing console queues (with the K Q command) can result in throwing away innocent and often important messages.

Message Flood Automation cannot handle DOM (Delete Operator Message) floods.