Presented at Quest Experience Week (QXW) - Database & Technology Day
Tired of continuously adjusting your performance monitoring and alerting rules? How about using machine learning to create the rules instead! Sure, we can train a supervised ML model to recognize patterns of poor performance. But what if your IT department is not ready to embrace it?
A novel solution is to train a supervised machine learning model to recognize patterns of poor performance, but then extract the rules in plain English and then manually enter them into your existing monitoring and alerting platform. Is this possible? Yes, it is! And, I will demonstrate how you can do this.
To ensure you can easily do everything I do, I will use industry-standard Python ML libraries, the industry-standard Jupyter notebook, AWR, and support ticket data. You can experiment with the demonstration materials “as is” and then use them to push your ML and rule creation knowledge deeper.