Insight Symptoms and Noise Reduction

In some cases, when searching for anomalies, a number of segments might be affected by the same anomaly.

For example, imagine that a metric is drifting for data originating from a specific device type. It might be the case that data originating this device type also usually originates from a specific browser. If a naive approach would have been taken, Mona would alert you twice - once on the device type and once on the browser type, which would be noisy and misleading.

However, Mona detects that the two segments are drifting due to the same anomaly, and alerts you just once, telling you what is the main way in which the anomaly can be viewed, and what are other “symptoms” of that same anomaly.