微服务故障检测研究综述  

A Review of Microservice Fault Detection

作　　者：王璐[1] 姜宇轩 李青山[1] 霍其恩 王展 谢生龙 歹杰 WANG Lu;JIANG Yu-Xuan;LI Qing-Shan;HUO Qi-En;WANG Zhan;XIE Sheng-Long;DAI Jie(School of Computer Science and Technology,XiDian University,Xi’an 710071)

机构地区：[1]西安电子科技大学计算机科学与技术学院,西安710071

出　　处：《计算机学报》2023年第11期2342-2369,共28页Chinese Journal of Computers

基　　金：国家自然科学基金青年科学基金项目(No.61902288);国家自然科学基金面上项目(No.61972300);陕西省自然科学基金青年项目(No.2020JQ-300)资助。

摘　　要：微服务架构的独立部署、去中心化等特点契合当前软件系统对适应变更、弹性部署、高扩展性等需求,使得该类架构得到了广泛应用.但这些特点同时也增加了软件配置、监控的难度,导致软件故障具有不确定性和不可控性,增加了对微服务软件开展故障检测的复杂性.然而,鉴于微服务故障检测对提升软件性能和可靠性的重要作用,目前也涌现出一系列挑战问题复杂性的研究工作,但目前缺乏对这些研究工作的系统性地梳理和总结,导致该领域的基础概念不清晰、研究方向界限易混淆和重叠,影响了该领域的研究进展.因此,本文对2009年至今共105篇研究工作开展了文献综述,不仅分析了关于微服务故障检测的相关工作,并讨论了单体架构、分布式架构等其他软件架构采用的故障检测方法应用于微服务架构的可能性.本文旨在帮助相关学者和从业者梳理该领域的最新研究进展,完善故障检测技术体系,以切实推进微服务架构的应用和发展.本文发现软件监控、故障诊断、故障预测、故障复现和测试是微服务故障检测领域的四个主流研究方向,因此本文进一步剖析了每个方向的研究现状与难点问题.并且,本文给出了一个建议的微服务故障检测基础框架,阐述了四个研究方向在框架内的业务范围和内在联系,并讨论了各方向的发展趋势,为该领域的后续发展提供未来方向.Because the characteristics of independent deployment and decentralization,the microservice architecture meets the requirements of current software systems for adaptation to changes,flexible deploy-ment,and high scalability,which make this kind of architecture widely used.However,these characteristics also increase the difficulty of software configuration and monitoring,leading to uncertain and uncontrolla-ble software faults.It increases the complexity of fault detection of microservice software systems.How-ever,in view of the important role of microservice fault detection in improving software performance and reliability,a series of research works challenging the complexity of problems have emerged.However,there is a lack of a systematic summary of these research works.It leads to unclear basic concepts,and confusing and overlapping research directions,which affects the research progress in this field.Therefore,this paper provides a literature review of 105 research works from 2009 to the present.It not only analyzes the related work on fault detection of microservice,but also discusses the possibility of applying fault detection methods adopted by other software architectures such as single architecture and distributed architecture to mi-croservice architecture. This paper aims to help relevant scholars and practitioners sort out the latest re-search progress in this field, improve the fault detection system, and effectively promote the application and development of microservice architecture. This paper finds that software monitoring, fault diagnosis, fault prediction, fault recurrence and testing are the four mainstream research directions of the fault detection of microservice. Therefore, this paper further analyzes the current research status and difficult issues in each direction. In addition, this paper gives a suggested basic framework for microservice fault detection. In ad-dition, this paper expounds on the business scope and internal connections of the four research directions within the framework, a

关 键 词：微服务 故障检测 软件监控 故障诊断 故障预测 故障测试 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]