Date of this Version
Krueger, C., Lohrman, B., Morrow, J., Nguyen, D., & Stapp, E.K. 2020. AIOps: Predictive Analytics with Mutual of Omaha Servers. Undergraduate Honors Thesis. University of Nebraska-Lincoln.
This project involved building and using an ARIMA model to process error logs from Mutual’s servers (a few days’ worth of data at a time for any one server or set of related servers) and predict whether the machine(s) error levels are going spike in the next 24 hours. If there’s a predicted spike, our system creates an alert that fits the same standards and model of Mutual’s other alert systems, providing machine names, predicted error times, and some baseline information about the potentially endangered machine so that engineers can mitigate quickly and effectively – stopping outages before they happen.