Stressperiod Option

From Messaging Server Technical Reference Wiki
Jump to: navigation, search

Introduced in release: 7.0.5

When a process such as the MMP becomes stressed due to high load or a denial of service attack, the stressperiod option controls how long (in seconds) the Watcher will consider that process stressed as well as how often that process will send a new stress notification to the Watcher. This allows msprobe to tell the difference between a wedged server (which should be restarted to improve the system) and a stressed process that is making forward progress but may have a response time larger than the msprobe timeout. Restarting such a process may reduce overall system performance as any disconnected clients are likely to reconnect to the system thus increasing the load. If a stressed process becomes wedged, then msprobe will be able to restart that process after the stressperiod expires.

This is presently only implemented by the MMP.

The default value is: 20

See also: