When you analyze the events that data observability generates for the anomalies observed in the characteristics of the data in your source system, you can resolve and reject the events that data observability generates. Data observability learns from your response so that it can provide more accurate events in the future.
You can perform the following operations for the events:
Resolve an event
If you are a stakeholder of a technical asset where the anomalies were observed, and if you find that the anomaly event is a correct observation, you can mark the event as valid. If a similar anomaly is detected again, data observability will create a new event to indicate the anomaly. To mark the event as valid, click Resolve on the event page.
For example, if data observability generates an event due to a sudden increase in the number of rows added to your source data from 10 to 10,000, verify whether the addition of 10,000 rows constitutes a valid anomaly that data observability identified. If you consider the anomaly identified by data observability to be valid, you can resolve it to address the generated event.
Reject an event
If you are a stakeholder of a technical asset where the anomalies were observed, and if you find that the anomaly event is an incorrect observation because you expected the change in your source data, you can mark the event as invalid. If a similar anomaly is detected again, data observability will not notify you. To mark the event as invalid, click Reject on the event page.
For example, if data observability generates an event due to a sudden increase in the number of rows added to your source data from 10 to 10,000, verify whether the addition of 10,000 rows constitutes an expected change. If you determine that the anomaly identified by data observability is an expected change, you can reject it to address the generated event.
The following image shows the Resolve and Reject operations for an event:
On an event page, you will learn about the type of anomaly that data observability has observed. Additionally, you will see a trend chart for percentage changes in metric values for the anomalies detected in your source data for the catalog source jobs that you run. The page also displays the stakeholders who can perform resolve and reject operations for events.
