7FRESH Agile Engineering Practices: Automation, CI/CD, and Metrics

This article details 7FRESH's agile engineering practices, emphasizing the critical role of engineering efficiency platforms in supporting agile implementation. The practices are structured around three core areas.

First, the article highlights the importance of automated testing platforms in agile delivery. 7FRESH's quality team has developed the "Falcon Intelligent Testing Platform" over several years. This platform enables teams to quickly improve test coverage and provides flexible support for test case management, reuse, and scenario configuration. The platform incorporates big data technology to analyze real production data and feed insights back into testing, combining test design with authentic data to make quality assurance more comprehensive and efficient, thereby preventing missed test cases and improving delivery efficiency.

Second, the article discusses continuous integration pipeline practices. 7FRESH has established a continuous integration pipeline where each stage can be automatically triggered by events and return results. The pipeline begins when developers submit code, triggering automatic code rule scanning that checks for potential defects, security risks, and code style issues. After code scanning, the system automatically packages and deploys features, then runs automated test scripts for system testing. Following automated testing, performance baseline load testing is automatically triggered.

The article emphasizes performance baseline load testing, explaining that instead of waiting for major events like 6.18 or 11.11 to conduct performance optimization, these activities can be integrated into daily work. For high-traffic applications, when updates occur, load testing is automatically triggered in the test environment. Although test environments differ from production, by keeping machines, pressure parameters, and network environments constant while only varying the code, the impact of code changes on system performance can be accurately observed and located. If performance issues arise, they can be immediately linked to recent deployments and optimized promptly.

Finally, the article covers the construction of a powerful agile metrics system. The need for this system arose from the urgent requirements of agile management functions. Current agile electronic tools focus more on the Kanban level with relatively weak statistical analysis capabilities, prompting the quality efficiency team to develop this system. The system focuses on measuring agile execution efficiency rather than specific product requirements or technical task descriptions. The system includes 30+ data parameters and 9 main indicators designed from multiple dimensions. To minimize team members' data update workload, the system automatically retrieves data from other systems whenever possible.

The system consists of three modules: iteration management, agile metrics, and data analysis. The first two modules are already in use, while the data analysis module is under construction. The system will display several agile metrics including product pre-launch ratio, requirement pre-launch ratio, average iteration requirement completion rate, average project delivery duration, average requirement delivery duration, and agile maturity assessment. The implementation of the metrics system can intuitively display agile efficiency and value, evolution trends, and expose problems and defects, helping agile coaches and Scrum Masters identify improvement points and solve problems in a targeted manner for continuous improvement.