How HAWK Turns Automated Testing into a Seamless DevOps Pipeline

From the "Testing Soul Three Questions" to Automation Challenges

Many testers feel undervalued, often blamed for issues, and see automation as a silver bullet that doesn’t always deliver. Common problems include scripts that exist only in quantity without results, lack of integration into the full development workflow, and limited use cases confined to regression testing.

Shortcomings of Traditional Solutions

Teams often focus on building an automation framework and consider the job done, yet the framework rarely lands in practice. Combining a framework with a test platform can help, but manual triggers limit frequent execution.

HAWK Platform: Framework + Platform + CI

The HAWK platform, created by the testing team, provides tools for interface testing, automated regression, mock services, and one‑click data construction.

1. Architecture Design

Automation scripts are version‑controlled in GitLab. Users configure case sets on HAWK, which can be triggered manually, by Git commits, internal applications, or AOS deployment hooks. After execution, HAWK collects results and generates reports. A built‑in “data factory” auto‑generates mock data for testing.

2. Function and Implementation

The automation process is likened to cooking: prepare ingredients (framework), process them (case‑set management), and add seasoning (data factory). This integrates automation throughout the development cycle, enhancing regression testing and data creation.

3. Preparing Ingredients – Automation Framework

The team uses SeLion for UI and App testing and a custom interface testing framework inspired by HttpRunner, employing YAML files and Test classes. The framework has three layers: case layer, framework layer, and engine layer (currently HTTP only). OKHTTP handles HTTP/HTTPS requests, and TestNG organizes test execution.

4. Semi‑finished Dish – Case‑Set Management

Case sets group automation scripts for easy execution and reporting. Users configure Git address, branch, and XML path or cases, then trigger runs via AOS WebHooks during code push, deployment, or release, achieving end‑to‑end automation.

5. Adding Seasoning – Data Factory

The data factory automates mock data creation for orders, products, and custom orders, allowing users to generate needed business data directly from the platform, dramatically reducing waiting time.

Results and Future Directions

Automation scripts and case‑set execution have increased regression efficiency by about 30%, and the data factory saves significant tester time. Future work includes integrating Sonar code scanning, API‑driven interface testing, expanding data factory types (SQL, MQ, etc.), deeper case‑set integration, and platformizing performance and App testing, all to support faster, higher‑quality delivery.