Design And Development Of Cloud Platform Monitoring System

In recent years,public cloud platforms and private cloud platforms have developed rapidly.Small and medium-sized enterprises have deployed their application services on public cloud platforms to save application deployment and server maintenance expenses.Large-scale enterprises and state-owned enterprises have also built private cloud platforms for data security and flexible deployment of services.The flexible resource allocation and flexible expansion of cloud services have solved the shortcomings of traditional service deployment for enterprises,while providing different services for users to choose.The rapid development of cloud platforms has brought challenges to cloud platform monitoring: with the diversification of cloud service forms,the monitoring dimensions must also be diversified,and the systems and applications in the cloud platform must be monitored simultaneously;the cloud platform should simplify user monitoring configuration to achieve automatic monitoring and deployment;more and more tests for cloud platform service performance require indicators to be supported independently of monitoring systems outside the cloud platform;at the same time,cloud platform providers use the platform to optimize the operating efficiency of the cloud platform research on monitoring data to redistribute cloud resources.From the above perspective,this paper first analyzes the monitoring system of the current cloud platform,and combines the current large independent distributed monitoring system to study the impact of monitoring frequency and the number of monitoring items on the monitoring of the cloud platform.Cloud platform monitoring solution.Using Zabbix as the monitoring component,the cloud platform monitoring system is designed and developed to realize multi-dimensional monitoring of the physical nodes,virtual nodes and applications in the virtual platform.In order to meet the monitoring data collection during performance testing and the monitoring data collection during daily cloud platform operation,this paper has designed two data transmission methods.Aiming at the four performance tests of cloud facility test platform dynamic expansion,load balancing,fault migration and node failure,the design and development of the computing module of the cloud platform monitoring system collects the calculation performance test indicators with high frequency and few monitoring items.After the screening and calculation,the real-time performance data is displayed for performance testers.For the data collection of the cloud platform during daily operation,the front and back-end modules of the cloud platform monitoring system are designed and developed to collect virtual machine and application data in a low-frequency,multi-monitoring manner,showing users the physical nodes and virtual in the cloud platform Basic information and performance data of the node.According to the above information,the data analysis of the virtual machine is carried out,and based on the analysis data,a resource redistribution scheme with memory as the main reference parameter is designed and implemented,and the virtual machine redistribution scheme is executed in a multithreaded manner to effectively reduce the allocation The possibility of failure.This article designs an automated configuration script to implement automated monitoring of Tomcat in a virtual machine.At the end of this article,the open source OpenStack cloud platform is selected as the cloud platform for monitoring and testing,and performance testing is performed on the cloud facility test platform.You can see the real-time data acquisition of various performance indicators.Tested in the cloud platform monitoring system,it can successfully monitor the monitoring data of physical nodes,virtual nodes and applications,and generate data analysis information,and finally tested and verified the resource redistribution scheme in OpenStack.