Threshold values, when breached, should optionally provide a declaration to make it possible to send SLAThresholdEvents if and only if the threshold os breached using a stepping approach.
Lets consider a watch that is used to record values. If the watch has a high threshold of 10, and values increase, every time the watch value is recorded with a value > 10, an SLAThresholdEvent will be produced.
So values of 11, 12, 13, 14, 15, etc .... will all produce an SLAThresholdEvent.
The idea of a step would allow the value to grow, but only be notified per "step". Lets examine a step of 2 with the above values:
Notifications would sent sent out for 11, 13 and 15
This allows for less thrashing on values that may be increasing (like CPU or memory utilization), still provides notification but in a more managed way.
Threshold values, when breached, should optionally provide a declaration to make it possible to send SLAThresholdEvents if and only if the threshold os breached using a stepping approach.
Lets consider a watch that is used to record values. If the watch has a high threshold of 10, and values increase, every time the watch value is recorded with a value > 10, an SLAThresholdEvent will be produced.
So values of 11, 12, 13, 14, 15, etc .... will all produce an SLAThresholdEvent.
The idea of a step would allow the value to grow, but only be notified per "step". Lets examine a step of 2 with the above values:
Notifications would sent sent out for 11, 13 and 15
This allows for less thrashing on values that may be increasing (like CPU or memory utilization), still provides notification but in a more managed way.