I am implementing this new diff control feature and I have some questions:
What is the difference between the metric deviation and the prediction? How can I determine an effective threshold for these two metrics?
Another question is regarding the Variance Intensity metric, what would be the best threshold for this metric?
I'm not sure if this helps specifically but I have a few links of information.
The first is background from the product manager at the time when we introduced the feature, to explain what we did and why:
There is also a video relating to it by the same product manager which may help to understand
Understanding Differential Analysis
This is a link on configuring the differential analysis control and does mention values to set for the alerts
This link talks about how the data can be viewed and can possibly be useful to determine some thresholds.
Monitor Application Stability Using Differential Analysis - CA Application Performance Management - 10.7 - CA Technologi…
If this doesn't help or if we can find more specific information, please let us know.
Thank you very much David,
So it would be correct to say that the variance intensity metric is based on the prediction and deviation metrics? If so, we must understand that these metrics (deviation and prediction), would not be used to create the alarms, only the metric variance intensity, correct?
You should absolutely use DA for alerts. As a matter of fact, DA is already being used by default for Team Center.
DA is a better option since it's dynamic baseline instead of hardcoding. We should have a KB out there that details how you should setup your alert thresholds to work with it.
Yes, I want to use the DA to make the alerts, this function is very useful, I just need to understand whether I use only the metric variance intensity or use the metrics of deviation and prediction.
You'll want to setup your alert on the variance metric for your particular metric grouping.