Search Effect Monitoring System Based On Web UI Automation Tools

With the development of mobile devices and 4G communication technology,the global Internet industry has grown explosively for several years.In the era of mobile Internet,many huge Internet companies have appeared around the world,such as Google and Amazon abroad,Alibaba and Tencent in China.Almost all head Internet companies have their own APPs with more than 100 million DAU.Facing such a huge number of active users,online service quality assurance and user experience are hot issues that Internet companies pay close attention to.This paper develops an online result monitoring system called Monitor for search engine based on Web UI automation tools.Monitor can not only check static pages,also known as interface monitoring,but can also support UI interactive monitoring.Monitor supports both clocked check and triggered check.Monitor can be divided into three parts:CASE framework,scheduling system and user platform.CASE usually refers to a monitoring unit,such as a python script.The Monitor CASE framework provides a configurable way to develop CASE.Test engineers only need to set a series of parameters to run a CASE.The CASE framework reads CASE configurations from database,then transfers configurations to specific monitoring strategies,and decides whether to alarm the relevant owners in charge of the CASE.The configurable CASE development model greatly reduces the cost of CASE development and improves the quality of CASEs.In order to ensure that CASE execution clusters of dozens of machines can execute thousands of CASEs stably and efficiently,Monitor has developed a scheduling system based on the Celery scheduling model.The Celery scheduling model is a basic producer-consumer model.The Monitor scheduling system adds a load balancing module to balance the load average of backend consumers.It can effectively lower down the load average of low-performance machines and improve the utilization rate of high-performance machines.The Monitor user platform is based on VUE and an open source framework called iView.The user platform supports online modification,creation and deletion of CASE. In order to help improving the quality of CASEs,test engineers can view detailed data of CASE operation on the user platform.CASE owners create processes to deal with alarms and trouble-shooting experience is accumulated on the user platform.Key performance data are analyzed and shown on the user platform,such as request QPS of online service from Monitor.Monitor now has thousands of CASEs and executes millions of CASEs every day,which has saved hundreds of millions PV losses for the company.Monitor has proved its ability to guarantee the quality of online service.