 |
To cut down on false positives of all above algorithms, we apply dual thresholds: a relative one, ツ rel as レ after over レ before exceeding ン rel, and an absolute one, ツ abs as the difference surpassing ン abs.
We use those two values in the post-detection phase to confirm the change is significant enough.
For example, CPU usage of a server changes from 3 percent to 6 percent, even the ツ rel is 200 percent, but it is not really a notable problem. While in contrast, if CPU usage change from 45 percent to 90%, still 200 percent in ツ rel, but that is an significant one
We will choose this value based on domain knowledge of metrics
|
IEEE/ICACT20260142 Slide.20
[Big Slide]