Open AccessImprove continuous integration and delivery methods by implementing a load testing stage
Author(s) -
A. V. Bereznyuk,
Anatoliy Makarenko,
Ganna Grynkevych,
S. G. Lazebnyy
Publication year - 2021
Publication title -
zv'âzok
Language(s) - English
Resource type - Journals
ISSN - 2412-9070
DOI - 10.31673/2412-9070.2021.013944
Subject(s) - computer science , benchmarking , process (computing) , reliability engineering , agile software development , integration testing , automation , software , systems engineering , software engineering , engineering , operating system , mechanical engineering , marketing , business
The contradiction between system stability and development speed development introduces new challenges to processes automatization. Significant increase of quality and speed of software development was reached due to agile method and continuous integration and continuous delivery (CICD) approach, which enabled rapid software changes by breaking development process into small iterative stages. New product version with new features incorporated is the result of each iteration. Then, there are automation tools typically used for testing and building. However, there is space for performance issues and delivery failures in this scheme. System scaling is generic approach for resolving performance problems, but the question how much the system should be scaled remains unanswered. Load testing procedure is commonly used but it is still not perfect, as testing environment is very different from production environment. Therefore, current paper is aimed at overcoming described difficulties by extending CICD approach with load testing automatization, system benchmarking, which allows system interruption minimization, and system scaling evaluation by scale factors calculation. Using these enrichments more precise testing result will be provided. We also suggest using such tools as Go-replay for traffic mirroring and Nagios for monitoring. Hence, this paper suggests enriching and increasing the efficiency of CICD approach. In addition, paper addresses methods of monitoring metrics collecting in centralized system. They will be used in further analysis and decision-making process regarding new product version in the automated mode.