Detecting ETL Duration Anomalies with Simple Linear Regression
BI Platform Architecture, Development & Administration
When an ETL is initially deployed it is expected to run for a certain amount of time based on testing during development. It does not always come to our attention that the duration the ETL takes to complete starts growing at an abnormal rate. Using Simple Linear Regression, we can detect the duration anomalies of ETLs. Will be going into detail on what Simple Linear Regression is and how you can use it to your advantage to detect ETL duration anomalies.