This disclosure describes techniques to identify anomalous patterns in customer workloads from database logs and to enable timely, corrective action that ensures uninterrupted operation of the database. Examples of anomalies include sudden increases (bursts) in the number of error messages written to a log file. An adaptive behavior norm is defined for each message type. Time instances or periods when the gap between messages of a given type in the database log deviate from the expected behavior norms are detected. A deviation from the behavior norm is a potential indicator of database problems. An anomaly detection tool outputs a ranked list of log statements exhibiting spikes of activity along with their time intervals that a database administrator (DBA) can examine to take corrective action. By automating anomaly detection, the valuable time of DBAs can be spent acting on issues rather than finding them.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.