Topic – Performance and Security Testing using AI and for AI applications

A Centralized Web based analysis tool(VPAT – VOIS Performance Analyzer tool) where response time, error, server, defect identifications, DB, Heap and Thread dump and Logs analysis can be done automatically and reports are generated within minutes and shared with the stakeholders automatically after the tests are completed. Tool is created using the strong features of python libraries, usage of robust machine Learning algorithms to predict and to provide Conclusions and to detect the anomalies, PowerBI for Graphs Display. The entire solution is wrapped using APIs and deployed on cloud machines making the solution/tool faster to access and easy to maintain. It’s user friendly and the user can download the final report in different format like PDF and send it automatically in email. For the AI/ML the methodology used was Exploratory Data Analysis followed by best Model Selection, Training the model and evaluating its performance, Hyperparameter tuning and deploying the model. Reaching an overall accuracy of around of 97% was achieved for True Positive scenario after various parameter were tuned in the model. The entire solution helps in preparing the results within minutes in a customizable format compared to the hours of manual efforts used earlier for results collations and analysis. Integrated with CICD pipeline will help in results to be automatically uploaded and test report sent to relevant stakeholders after the test completions without manual intervention. The tool is developed in-house without any eternal sources involvement.