Abhijit Bangde

Topic: Speed Up Your Regression Testing Cycles with Data Analytics

Abstract: Use Case:
In today’s fast-paced dev environment where the code base grows rapidly and dev teams constantly update and modify their code with multiple incremental changes and feature additions, it is difficult to keep track of all the changes. Although dev teams are required to write tests for all these changes or perform regressions to check the negative impact of these changes, the lack of visibility means that teams have trouble identifying untested code. They cannot tell what part of the code was tested, or even how much of the application was ever tested. As such, no one can tell if the present build is better or worse off than previous builds.

This leads to bad release decisions as a result of inefficient data and insufficient insights into the areas of code that matter the most. In such environments, many code changes eventually reach production without being tested, resulting in the release of bug-filled, sub-optimal software. Such changes are referred to as “quality risks” and they have a huge impact on release quality. Due to the inherent difficulty in generating, collating and analysing all the necessary data required to accurately determine release readiness (especially in real-time), dev teams with no visibility or insights into the changes they’ve made or tests they’ve executed during sprint runs will experience decreased software quality.

Solution:
Coverage Analytics (Test coverage analytics using code coverage) was built to help dev teams dramatically reduce the time for test cycles and CI time (for both manual and automated tests) by providing immediate insights into the tests impacted by code changes. It reduces the total number of tests needed to run by selecting only the tests associated with the latest round of code changes. This helps software development teams focus their testing efforts, thus accelerating the development pipeline by speeding up the CI process.

Through the intelligent combination of code analysis, machine learning, and data mining techniques, Coverage Analytics recommends and executes the most impactful set of tests for build runs. Such data-driven decisions are key to increasing release velocity in fast-paced development environments and makes for more effective sprint planning. Coverage Analytics uses data from various sources within the software development pipeline to enable dev teams to prioritize and execute regression tests based on their efficacy in detecting bugs. This makes for shorter regression tests times, facilitates faster feedback on code changes that have caused test failures in the CI process and cuts down overall continuous integration time.