Do you want to rely on manual intervention to fix your application if something goes wrong? In this deep-dive session you will learn how Priceline uses machine learning to find outliers and anomalies in various data sets, including but not limited to bookings, search patterns, changes in logging patterns, etc. You will learn how we used machine learning combined with predictive analytics to solve variety of use cases. For example, we collect Kafka offset data, which is sending data to their respective syncs. We also monitor to see if the traffic is receded or data consumption has increased or decreased unexpectedly. We will show how different stages of application states are controlled with the use of data and alerts, like disabling the app and enabling it according to the data. We also will show you how Priceline deals with brownouts, the gradual degradation of volumes by using machine learning over long periods, using different self healing techniques and custom apps.

Speaker(s)
Mukund Murthy, Software Engineer, Priceline.com
Pranav Nandedkar, software engineer, Priceline.com

Slides PDF link - https://conf.splunk.com/files/2019/slides/FN1916.pdf?podcast=1576909588