SourceWarp: A scalable, SCM-driven testing and benchmarking approach to support data-driven and agile decision making for CI/CD tools and DevOps platforms

Abstract

The rising popularity and adoption of source-code management systems in combination with Continuous Integration and Continuous Delivery (CI/CD) processes have contributed to the adoption of agile software development with short release and feedback cycles between software producers and their customers. DevOps platforms streamline and enhance automation around source-code management systems by providing a uniform interface for managing all the aspects of the software development lifecycle starting from software development until software deployment and by integrating and orchestrating various tools that provide automation around software development processes such as automated bug detection, security testing, dependency scanning, etc..Applying changes to the DevOps platform or to one of the integrated tools without providing data regarding its real world impact increases the risk of having to remove/revert the change. This could lead to service disruption or loss of confidence in the platform if it does not perform as expected. In addition, integrating alpha or beta features, which may not meet the robustness of a finalised feature, may pose security or stability risks to the entire platform. Hence, short release cycles require testing and benchmarking approaches that make it possible to prototype, test, and benchmark ideas quickly and at scale to support Data-Driven Decision Making, with respect to the features that are about to be integrated into the platform.In this paper, we propose a scalable testing and benchmarking approach called SourceWarp that is targeted towards DevOps platforms and supports both testing and benchmarking in a cost effective and reproducible manner. We have implemented the proposed approach in the publicly available SourceWarp tool which we have evaluated in the context of a real-world industrial case-study. We successfully applied SourceWarp to test and benchmark a newly developed feature at GitLab which has been successfully integrated into the product. In the case study we demonstrate that SourceWarp is scalable and highly effective in supporting agile Data-Driven Decision Making by providing automation for testing and benchmarking proof-of-concept ideas for CI/CD tools, chained CI/CD tools (also referred to as pipeline), for the DevOps platform or a combination of them without having to deploy features to the staging or production environments.