Oracle Corporation
Detecting datacenter mass outage with near real-time/offline using ML models

Abstract:

The present embodiments relate to data center outage detection and alert generation. An outage detection service as described herein can process near real-time data from various sources in a datacenter and process the data using a model to determine one or more projected sources of a detected outage. The model as described herein can include one or more machine learning models incorporating a series of rules to process near-real time data and offline data and determine one or more projected sources of an outage. An alert message can be generated to provide the projected sources of the outage and other data relevant to the outage.