PASER:加性多维KPI异常根因定位模型
作者:
作者单位:

作者简介:

通讯作者:

靖宇涵,E-mail:weatherjyh@163.com

基金项目:

国家重点研发计划(2018YFB1800502);国家自然科学基金(61671079,61771068);北京市自然科学基金(4182041)


PASER: Root Cause Location Model for Additive Multidimensional KPIs
Author:
Affiliation:

Fund Project:

The National Key R&D Program of China (2018YFB1800502); The National Natural Science Foundation of China under Grants (61671079,61771068); The Beijing Municipal Natural Science Foundation under Grant (4182041)

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
    摘要:

    利用多维属性关键性能指标(KPI,Key Performance Indicators)的可加性特征能够实现对大型互联网服务故障的根因定位.由一项或多项异常根因导致的KPI数据变化会导致大量相关KPI数据值的变化.本文提出了一种基于异常相似性评估和影响力因子的剪枝搜索异常定位模型(Pruning Search Model based on Anomaly Similarity and Effectiveness Factor for Root Cause Location,PASER),该模型以多维KPI异常传播模型为基础,提出了衡量候选集合成为根因可能性的异常潜在分数评估方案,基于影响力的逐层剪枝搜索算法将异常根因的定位时间降低到了平均约5.3秒.此外,本文针对异常根因定位中所使用的时间序列预测算法的准确性和时效性也进行了对比实验,PASER模型在所使用的数据集上的定位表现达到了0.99的F-score.

    Abstract:

    Additivity of multidimensional KPIs (key performance indicators) was used to achieve root cause location for large-scale Internet services. The anomaly caused by one or more root causes usually results in the change of a large number of relevant KPIs. A pruning search model based on anomaly similarity and effectiveness factor for root cause location (PASER) was proposed, which indicated the probability of candidate set becoming root cause using potential score based on the anomaly propagation model of multi-dimensional KPI. The pruning search algorithm used in PASER also managed to reduce the location time to about 5.3 seconds on average. In addition, the selection of time series prediction algorithm was also discussed. PASER had finally achieved a performance of 0.99 F-score on the experimental dataset.

    参考文献
    相似文献
    引证文献
引用本文

靖宇涵,何波,张凌昕,李天星,王敬宇,刘聪. PASER:加性多维KPI异常根因定位模型.软件学报,,():0

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
历史
  • 收稿日期:2020-08-01
  • 最后修改日期:2020-10-15
  • 录用日期:
  • 在线发布日期: 2021-01-15
您是第位访问者
版权所有:中国科学院软件研究所 京ICP备05046678号-3
地址:北京市海淀区中关村南四街4号,邮政编码:100190
电话:010-62562563 传真:010-62562533 Email:jos@iscas.ac.cn
技术支持:北京勤云科技发展有限公司

京公网安备 11040202500063号