Systems Monitoring via Subspace Tracking Algorithms
October 21, 2016
12:20 pm - 12:40 pm
GRB 340 A
Track:
Data Science
Type: Presentation
Level:
Beginner/Intermediate
Detecting unexpected changes in resource consumption is vital to any organization that utilizes a distributed cloud computing environment. Automated systems monitoring techniques often model each time series individually and/or offline. In a novel approach, we discuss how to exploit the redundancy in systems metrics to detect changes in real time without explicitly modeling seasonality or noise.
Speaker(s)
, Data Scientist, Twitter, Inc. |