Service monitoring and failure fast positioning method adopting self-detection mechanism

Abstract

The present invention discloses a service monitoring and fault quick locating method by adopting a self-check mechanism. the method is characterized in that it comprises steps of making the hybrid codes of self-check service request identification information and self-check service request start time information form a virtual user key ID, generating a self-check service request through the virtual user key ID, a service supporting system processing the self-check service request, collecting system feedback results triggered by the self-check service request, at the same time, searching in logs of each service processing link of the service supporting system by using the virtual user key ID as a searching key word, finally, according to the system feedback result and the searching result, comparing related situations of normal service requests, judging the health state of the system and locating the system fault points. According to the present invention, service monitoring and fault quick-positioning of the service supporting system can be realized on the basis of least alteration of the original system, therefore, outlay cost of a corporation can be reduced and the method has a high practicability.

Description

technical field [0001] The invention relates to the monitoring, analysis and troubleshooting of a business support system, in particular to a business monitoring and fault quick location method using a self-check mechanism. Background technique [0002] With the development of computer technology, the business of each company began to be gradually integrated with computer software and hardware. According to the company's business model, the computer system was the core, combined with other software and hardware facilities, and customized to form the company's business support system. Such as the billing system of the communication industry, the transaction management system of the bank, the transaction and payment system of the website, etc., are all typical business support systems. [0003] In the business support system, some key businesses need to run stably for a long time. In the operation of the company, the sooner the faults of these business systems are found and th...

AI Technical Summary

Problems solved by technology

[0006] The system monitoring method needs to set up detection points in the system, use special detection instruments or write special detection programs. During implementation, it is necessary to modify all aspects of the system, which is costly; special detection instruments or detection programs are easy to affect the monitored system , when there is a problem with the detection instrument or the detection program itself, it may cause false alarms and even destroy the monitored system

In reality, there may be many business programs running on the same host, and these businesses may be independent of each other, but any problem with any business program may cause abnormalities detected by monitoring checkpoints (such as the usual cpu usage, memory usage, etc.) and other monitoring indicators are abnormal), and it is difficult to locate which business program has a problem from these abnormal indicators, that is, the system monitoring method is difficult to monitor by business

[0007] Adopt the idea of ​​system monitoring, improve the monitoring method, set checkpoints according to the business, such as checking whether the business log is updated, whether the key process of the business exists, whether the business volume reaches the threshold in a certain period of time, etc., can be realized to a certain extent Monitoring of business, but most of these inspection methods can only check the state quantity (such as log update / not updated, process existence / absence, exceeding threshold / not exceeding threshold), and these indicators are different from normal business volume The changes are closely related, so it is difficult to distinguish and judge whether the detected state changes are caused by system abnormalities or business volume changes. It is difficult to guarantee the accuracy of monitoring, and it is prone to false alarms and unreported failures.

The biggest problem with this method is that in order to locate faults, every link of the business process must be monitored as much as possible, and the construction cost of the monitoring system is very high.

[0008] The fault dictionary method requires careful consideration of possible faults during system design, and a large number of tests in advance to simulate system faults and compile a fault dictionary, so there is also a problem of high cost; the fault dictionary is often not available according to the fault phenomenon. To locate the fault, it is necessary to use various auxiliary means such as measurement. When the auxiliary means such as measurement are needed, the fault point cannot be accurately located, and the auxiliary means such as measurement often increase the cost of system troubleshooting and may have adverse effects on the system; Another big problem with the dictionary method is that it cannot detect and locate unknown faults. This method is no longer effective when the fault phenomenon exceeds the range of the fault dictionary.

[0009] More importantly, there are many problems in the implementation of the system monitoring and fault dictionary methods for the existing business support systems. Because of the market situation at the beginning, the measurement points may not be considered in the design of the business support system, and the new measurement Due to the limitation of the development progress, the establishment of fault dictionary and the coding index of the fault dictionary (such as fault codes) may not be considered. Re-establishing fault codes may also require more system transformation , there are greater difficulties in the compilation and use of fault dictionaries