基于深度神经网络的异常检测算法在国网上海公司AIOps中的应用

Application of Deep Neural Network Based Anomaly Detection Algorithm in the AIOps of State Grid Shanghai Company

AIOps是IT运维中一个极具革命性的概念,它以大数据和算法为基础,完善现在的运维体系。实现AIOps的体系构建最重要的是对监控采集的指标数据进行趋势预测以及异常检测。文章分别针对这两方面提出相应的算法,为实现AIOps更高级的功能奠定基础。

AIOps is a revolutionary concept in IT operation and maintenance.It subverts or perfects the current operation and maintenance system from the perspective of big data and algorithm.The most important thing for realizing the system construction of AIOps is to forecast the trend of the index data collected by monitoring and detect the anomaly.This paper proposes corresponding algorithms for these two aspects to lay the foundation for the realization of more advanced functions of AIOps.