Anomaly Detection in Cloud Components

TL;DR

This paper presents a method for early anomaly detection in cloud components using a Gated-Recurrent-Unit autoencoder, aiming to improve cloud service reliability by analyzing resource utilization metrics.

Contribution

It introduces a novel application of Gated-Recurrent-Unit autoencoders with likelihood functions for anomaly detection in multi-dimensional time series of cloud metrics.

Findings

High detection performance achieved

Effective in multi-dimensional time series

Potential for early failure detection

Abstract

Cloud platforms, under the hood, consist of a complex inter-connected stack of hardware and software components. Each of these components can fail which may lead to an outage. Our goal is to improve the quality of Cloud services through early detection of such failures by analyzing resource utilization metrics. We tested Gated-Recurrent-Unit-based autoencoder with a likelihood function to detect anomalies in various multi-dimensional time series and achieved high performance.