DOP 180: What is AIOps?
DevOps Paradox
00:00
The Trade-Off Between Observed Data and Predictive Modeling
When collecting data samples, one week's worth of data was useful. There was still a strong likelihood of false positives based on the fact that four minutes past 11 wasn't exactly the same as last week. Two weeks of data considerably improved the results. Three weeks of data improved it, but not the same level of improvement you got after the first week.
Transcript
Play full episode
