Design And Implementation Of Multi-tenant Resource Monitoring System For SaaS Platform Of Automobile And Parts Industry

As SaaS(software as a service)model develops more and more maturely and becomes successful commercially,many enterprises are joining the team of SaaS.After renting the services by choosing packages designed by the SaaS providers,tenants can obtain one-stop business solutions.At the same time SaaS service providers not only reduce the cost,but also expand the size of the economy by sharing and reusing of SaaS service.The SaaS platform of automobile and parts industry provides tenants with selling passenger cars,commercial vehicles,customer management,and other services,and design a wide variety of rental package including network traffic,storage resources.Tenants use the services providing by the platform after they rent some package.When the tenants are using the services what they rents,the platform needs to dynamically monitor the amount of network traffic and the amount of storage resources that they are using,and can master the usage of its resources in real time.In addition,because the platform database has stored the business data resources of tenants,the platform also need the main operation of online monitoring platform database in order to ensure that the tenant can use normally business data resources.Based on the monitoring requirements of the SaaS platform of automobile and parts industry,this thesis designs and implements the multi-tenant resource monitoring system.Mainly including:(1)Analyzing the requirements of the platform in depth,this thesis realizes the dynamic collecting and processing of the tenants' network traffic by the flow acquisition technology based on Jpcap and the tenant recognition scheme and it uses the AOP to collect and process dynamicly the change of the tenants' storage resource.As a result,this thesis can accurately record the amount of network traffic and storage resource of each tenant.(2)The monitoring system realizes monitoring the main running status of the database on the platform,including QPS,TPS,the number of DML,and thread status.The monitoring system realizes analysing the slow query log and monitoring of user permissions.(3)For the different monitoring targets,the monitoring system formulates different monitoring rules and dispatching strategies,and realizes the real-time dynamicly monitoring alarm.To a certain extent,the monitoring system provides a warning management for platform administrator and safeguardes the healthy and stable platform to run.(4)By analyzing and comparing the web message push technology,the real-time data push technology based on WebSocket is used to dynamically draw the monitoring chart and provide real-time information of the monitoring targets by monitored for the administrator.(5)Through the statistical analysis of the historical data of the monitoring system,it provides administrators with various statistical reports and provides some help for their decision-making.(6)By adopting the component management based on Io C container,the monitoring system achieves the goal of modularization,so that the monitoring system has a low coupling,high maintenance,high extensibility,etc,and it provides a good foundation for new monitoring demands and maintenance and upgrading of system.