A method of software runtime performance analysis

Abstract

The embodiment of the invention provides an analysis method of software operating period performance. The analysis method comprises the following steps: starting a software system to be analyzed, wherein the software system to be analyzed is provided with a probe in advance; installing the probe into the software system through a bytecode; through the probe, collecting the performance index of the software operating period; and according to the performance index, analyzing the software operating period performance. Through a dynamic plugin technology, a performance probe is provided by aiming at the performance index which needs to be monitored by the software, each meaningful performance index is collected, filtered and stored in a software operation process through a way of software operating period configuration, and a software development process is not invaded. During the operating period, the performance index which is configured at present permits to be dynamically added, revised and deleted so as to change the collected performance index, the performance probe supports the sampling and the filtering of the performance index, ineffective data is reduced, and data accuracy is improved.

Description

technical field [0001] This application relates to the field of computer technology, and in particular relates to a method for analyzing software runtime performance. Background technique [0002] Software operating performance is one of the most important non-functional characteristics of software, and software performance optimization is usually based on software performance quantification and analysis results. [0003] Most of the general monitoring tools or methods during the software running period are comprehensive monitoring or scanning methods. They monitor every detail of the software running period and have a great impact on the performance of the running period. After deployment, it will affect the software performance, so they cannot be used in the production environment. [0004] Another method of quantitative analysis of software performance is based on logs, that is, the log output is fixed during the development period, and the output logs are analyzed during...

AI Technical Summary

Problems solved by technology

[0003] Most of the general monitoring tools or methods during the software running period are comprehensive monitoring or scanning methods. They monitor every detail of the software running period and have a great impact on the performance of the running period. After deployment, it will affect the software performance, so it cannot be used in the production environment.

[0004] Another method of quantitative analysis of software performance is based on logs, that is, log output is performed by solidifying the code during the development period, and the output log is analyzed during the operation period. This method is very intrusive to software development and coding. Change the monitoring target and location during the runtime. In addition, due to the limitation of the log system, the output often exists in the text system, which makes it impossible to directly use the log output for performance analysis. The process of conversion and extraction is also required, and the cost of subsequent analysis data is too high.

Images