互联网企业线上服务实时监控系统研究

发布时间：2018-01-30 04:28

本文关键词： 监控系统健康度评价业务监控低入侵　出处：《大连海事大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着互联网的发展,智能设备的普及,更多的业务被迁移到网络,人们的生活和工作也日渐离不开网络服务。互联网行业竞争激烈,确保线上服务持续稳定的运行成为企业关注重点。传统监控工具主要监控IT资源层性能,无法对业务功能实施监控。各工具的监控功能相互独立,企业需使用多个监控工具达到全面监控线上服务的效果。管理人员根据单一故障报警无法掌握服务整体的运行状态。企业各部门为保障线上服务运行质量,独立开发、部署一套监控系统。对企业而言,这种方式是不经济的。企业迫切需要一套企业级通用的监控系统满足各部门的监控需求。本文在充分研究现有监控系统的前提下,结合互联网企业对线上服务的监控需求,给出线上服务监控系统架构,架构由数据采集层、数据层、应用层和展示层组成。重点研究线上服务低入侵的监控数据采集技术来满足各部门的监控需求。研究业务功能监控弥补传统监控工具对功能层监控的缺失。结合资源层性能监控和业务功能监控全面监控线上服务,使用户可以更全面的掌握线上服务的运行状态。研究模糊综合评价方法,基于该方法给出线上服务健康度评价算法。算法以监控系统采集的各性能数据为基础,通过建立线上服务、IT资源及性能指标之间的关联,从而计算出服务的健康度。服务健康度对线上服务的管理及故障的排查具有指导意义。给出分布式数据存储方案解决海量监控数据存储问题。基于上述理论研究,设计并实现线上服务实时监控系统,给出系统功能模块设计,实现系统核心模块,包括监控项配置模块、监控项调度模块、业务功能监控模块、服务健康度评价模块。本文研究的监控系统企业级通用,全面监控线上服务运行状态。目前,系统已投入使用,从使用效果来看系统降低了公司对线上服务的监控成本,减轻用户工作压力,具有实际意义。
[Abstract]:With the development of the Internet and the popularity of smart devices, more business has been transferred to the network, people's lives and work are increasingly inseparable from the network services. The Internet industry competition is fierce. The traditional monitoring tools mainly monitor the performance of IT resource layer and can not monitor the business function. The monitoring function of each tool is independent of each other. Enterprises need to use a number of monitoring tools to achieve the effect of comprehensive monitoring of online services. Managers can not grasp the overall operating state of the service according to a single fault alarm. All departments of the enterprise ensure the quality of service on the line. Independently develop and deploy a monitoring system. For the enterprise. This way is not economical. Enterprises urgently need a set of enterprise-level general monitoring system to meet the monitoring needs of various departments. This paper fully studies the existing monitoring system. According to the monitoring demand of Internet enterprises, the architecture of on-line service monitoring system is given. The architecture is composed of data collection layer and data layer. The application layer and display layer are composed. The monitoring data collection technology of low intrusion service on line is studied emphatically to meet the monitoring requirements of various departments. The research on business function monitoring makes up for the lack of traditional monitoring tools to function layer monitoring. Resource layer performance monitoring and business function monitoring the overall monitoring of online services. The fuzzy comprehensive evaluation method is studied. Based on the fuzzy comprehensive evaluation method, the algorithm for evaluating the health degree of the online service is given. The algorithm is based on the performance data collected by the monitoring system. By establishing relationships between online service IT resources and performance metrics. Thus, the health degree of the service is calculated. The health degree of the service is of guiding significance to the management of the online service and the troubleshooting of the fault. A distributed data storage scheme is presented to solve the problem of massive monitoring data storage. . Design and implement on-line service real-time monitoring system, give the system function module design, realize the core module of the system, including monitoring item configuration module, monitoring item scheduling module, business function monitoring module. Service health evaluation module. This paper studies the monitoring system of enterprise general, the overall monitoring of the service running status on the line. At present, the system has been put into use. From the effect of use, the system can reduce the monitoring cost of the online service and reduce the working pressure of the users, which is of practical significance.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP277

【相似文献】