SYSTEMS AND METHODS TO IDENTIFY AND CLASSIFY PERFORMANCE BOTTLENECKS IN CLOUD BASED APPLICATIONS

摘要

Described herein are systems and methods of identifying and classifying performance bottlenecks for web applications. Such systems and methods use classification and analysis of performance testing data and data instrumentation via arithmetic and/or machine learning. Data is integrated from different sources including system data, historical and real time sources. Performance variations are analyzed as load changes and the impact of these variations on different sectors of the Application stack are analyzed. Bottlenecks are identified and classified based on the sector in the software stack and recommendations for optimization of an Application under Test are presented to address the bottlenecks are presented.

主权项

1. A system for identification and classification of performance bottlenecks in web applications comprising:
a test execution module to generate synthetic traffic, run a performance test on an Application under Test and collect raw data from execution of the performance test; at least one data instrumentation system which collects system data; a pre-processing module which groups the raw data from a data instrumentation system, the performance test, and historical data into functional groups based on mapping schema generated by the at least one data instrumentation system and maps them to a metric; an analytics processing module which compares the functional groups assembled by the pre-processing module and identifies a threshold of a relevant metric associated with a performance bottleneck; a plurality of classification modules which identify a sector in the Application under Test that are more highly impacted by changes in load; and a display mechanism that provides a graphical representation of the performance bottlenecks and metrics and recommendations for optimizing performance of the Application under Test.