面向集群的系统性能可视化监控平台

Visualization Platform for Cluster-oriented System Performance Monitoring

伴随着计算机技术的不断发展,对于系统服务稳定性的要求显得尤为重要,集群的监控必不可少。设计并实现一个面向Linux集群的系统性能监控管理器,以检测集群运行状态。该监控管理器采用C/S(客户端/服务器)模式,通过多线程技术完成数据采集处理、缓存技术实现数据查询和HTTP协议完成数据传输,并以可视化图形界面实时显示系统当前状态和对历史数据的查看,同时当系统某一参数超出正常指标,可远程处理瓶颈,以确保系统服务的稳定性。

With the continuous development of the computer field, the stability of the system service is particularly important, and the cluster monitoring is essential. This article studies the design of a Linux cluster to detect the cluster service running state for the system performance monitoring manager. The monitoring manager uses C / S （client / server） mode, multi-thread technology is used to complete data acquisition and processing, cache technology to achieve data query and HTTP protocol, data transmission is then completed. Visual graphical interface is designed to real-timely display system current status and history data. While a parameter is abnormal, manager can remote process bottlenecks to ensure the stability of system services.