Design And Implementation Of The Cloud Monitoring System Based On Zabbix

With the rise and development of the software-as-a-service concept in cloud computing,the service-oriented system is accepted by more and more enterprises and users.Enterprises benefit from the rational utilization of the resources and users benefit from the high quality of the service.The topic of the thesis comes from the project of the community cloud main station monitoring system.When a problem occurs at the spot of the traditional master station,the user needs to feedback the problem and wait for the operation and maintenance personnel to solve it remotely.When the user finds the problem,the abnormality has caused losses,which wastes manpower,financial resources and time,which seriously affecting the user experience and operation and maintenance efficiency.Through the centralized management of cloud computing resources,the real-time monitoring,diagnosis and alarm of metrics,the system solves the problem of low efficiency and resource decentralized management in the traditional operation and maintenance mode.The traditional alarm mode uses a fixed threshold alarm policy with strong subjectivity.Therefore,the setting of the flexible alarm threshold and accurate judgment of abnormal metrics have become a difficult problem in the project.This paper designs and implements a cloud monitoring system by researching and analyzing Zabbix and Prophet based on the current business requirements of the company with the internal and external situation of the monitoring system and carries out the following works:1)Zabbix is used as a data collection tool to collect server,network and application performance metrics in clusters.Visualizing the real-time data of the monitoring metrics on the web page with the back-end interfaces provided by Zabbix;2)Prophet is used to predict the time series metrics of the server,dynamically adjust the alarm threshold and perform alarms based on the predicted results by Prophet;3)The alarm information is sent by email and SMS based on the diagnosis result to find problems timely and reduce losses.4)Metric diagnosis configuration,abnormal alarm configuration,abnormal informationmanagement,and alarm information management are also provided;5)Tested metric collection,metric diagnosis,and exception alarm in system functional requirements and system performance in system non-functional requirements.The monitoring system has been fully tested and put into use in the company,which not only can meet the requirements of the cloud environment,but also improve the operation and maintenance efficiency and relieve the operation and maintenance personnel.It basically meets the needs of daily operation and maintenance work,and is of great significance and far-reaching significance for the development of the community cloud master station system and electricity information collection system.