Does anyone know any work that justifies the identification / selection / definition of values for setting limits / quality thresholds?
For example:
Metric:% CPU usage
lower threshold - below 70% - deallocates
upper limit - above 80% - allocates
between lower and upper threshold - maintains allocation
I need references for these thresholds.
Well, i need a way to justify values used in thresholds.
In many related work what I found was:
-use of fixed values without justifying why
-values dynamically adjusted
Often these values are defined by application type, but I miss work that explain how to identify these values in a more applied way.
For example: I use in my workloads CPU intensive applications, such as math benchmarks. To use a mechanism that provides elasticity based on thresholds, it is interesting that I know how to define the range of values.
- Badges
- Science method