You want to detect anomalies in your ordering process to proactively remove obstacles, so you used the Order (COUNT) as the metric for anomaly detection.
After a period of steady traffic that generated a steady number of orders (say 1000 per day), something happened.
Scenario A: Anomaly detected
Question: There was a technical error (an event that detects the error is defined in the system) that was presented to 200 customers. This error caused 100 orders to fail, so there was a drop in the number of orders by about 10%. Will this issue be detected as an anomaly?
Answer: The system should detect this error as one of the top correlated factors if the error code is in the dimension that was pointed to as a correlated factor.
Scenario B: Anomaly not detected
Question: There was a one-day flash sale organized by the marketing department. There was more traffic on the website, and there were 100 additional orders placed on that day. Will this issue be detected as an anomaly?
Answer: If there is no Tealeaf event that relate to the marketing activity, no contributing result is shown. There are no Tealeaf events that detect external marketing activities, so none of the other metrics contributed to the anomaly.