The Design And Implementation Of General Transaction Monitor And Alert System Based On Groovy

With the development of E-Commerce on the Internet,online transactions have been closely linked with people's life.In order to monitor the status of the running transaction system or know the trading results,all companies and teams need to design and develop a monitor and alert system.But how to improve the general-utility and usability of the monitor and alert system and how to realize flexibility of the alert rules?These are the problems this paper wants to resolve.This paper describes the research and development of a set of general transaction monitor and alert system base on groovy,which can be applied to most of the transaction system without developing another separate system,it can monitor the statistical results of the transaction and send warnings by the alert rules.This paper studies the current development of monitor and alert systems,note others system's architecture and technology.In general,the whole system is divided into four parts,data acquisition,calculation and statistics,data monitoring and elastic alert.Data acquisition is responsible for collecting the raw data from the original transaction system and storing the data into the monitor and alert system.The calculation and statistics part is in charge of calculating the raw data into the statistical indicator data and storing it.The data monitoring part takes the job of offering services related to data monitoring.While the elastic alert part do the task of sending the warnings about abnormal phenomena according to the rules set by users.In this paper,it describes a general transaction monitor and alert system,which finds out a solusion with four aspects as follows.1.A data acquisition method based on the log file.The solution of data acquisition defines a log format and collects raw data from the original transaction system by analyzing the log file.The solution applies to most of the transaction systems and requires the minimum change or no change on that system.2.A statistical calculation based on SQL.The solution uses statistical function of the database and calculates the indicator data according to the indicator information.3.A management of alert rules base on groovy.The management of alert rules uses the groovy script as the alert rules because it can seamlessly and transparently integrate and interoperate with Java,support runtime and compile-time meta-programming which greatly improves the scalability of the system and makes it possible to let the system run all the time.4.A RESTful architecture of related management services.The system realizes many management services related to monitoring and alerting with RESTful architecture style using framework of Spring,Jersey and Mybatis.The management services include log model management,indicator management,alert management,role management and subscription management.The monitor and alert system described by this paper has a good universality that applies to most of the transaction systems and has a good extensibility that it can calculate many types of indicator data and set various kinds of alert rules made by users.