Automated Activity Regression and Configuration Validation for Social Live Streaming

Based on the special characteristics of social live‑streaming activities, this article presents an automated regression and configuration‑validation platform that improves efficiency and ensures the quality of online events.

Background

Social live‑streaming activities account for a large share of the business, featuring short cycles, high frequency, and diverse gameplay. With limited manpower, quality assurance faces two main challenges:

High regression manpower : After a feature is developed, each subsequent activation of the activity requires a full regression, leading to excessive manual effort.

Configuration checking errors : Complex gameplay results in massive operational configuration documents. Operators must locate relevant information manually, and developers, QA, and planners must perform time‑consuming manual checks of online configurations.

To address these issues, separate automated workflows for activity regression and configuration validation are implemented.

Activity Regression Automation

Operational activities have relatively fixed scenarios, but manual regression is costly and vulnerable to staff turnover. The existing GoAPI platform presents difficulties: data construction, complex scenario cases, limited validation flexibility, and lack of comprehensive reports.

The solution combines platform scheduling with test‑script execution, covering data generation, scenario execution, result validation, and visual report generation. The overall flow includes a task management system that creates, executes, and visualizes reports; Jenkins‑driven script execution for each activity; asynchronous result reporting; and grouped visual reports.

Activity Configuration Validation Automation

In large‑scale events, multiple sessions with different gameplay types are planned by various operators. The current manual process involves documenting configurations in shared files, entering data into backend consoles, and performing manual online checks, leading to:

Unstandardized, voluminous configuration documents that are hard to navigate.

Scattered online configurations that are prone to omission during manual checks.

The proposed solution standardizes configurations into four modules—ranking, task, lottery, and redemption—and builds an end‑to‑end validation pipeline consisting of:

Document : An online mind‑map style document that records all configuration data with outline and graph views.

Template : A unified template that extracts verifiable attributes from raw configuration data.

Validator : A validator that compares online configuration against the template and outputs results.

The operational flow mirrors a mind‑map interface: a “document” aggregates all activity configurations; nodes can be added, edited, or deleted; double‑clicking a node opens its configuration panel (e.g., task, ranking, lottery, redemption). After triggering validation, the backend asynchronously classifies results into four categories—online consistent, missing, extra, or differing—and aggregates them into hierarchical reports by sub‑activity, result type, and specific configuration item.

Summary

Regression automation integrates task management, Jenkins scheduling, script execution, and visual reporting, reducing manual verification time for leaderboard and task modules from 1 day to 0.2 day and enhancing stability.

Configuration‑validation automation provides document management, query capability, and automated diffing, cutting alignment effort from 0.5 day per participant to 0.1 day, especially beneficial for large events.

The article is published by NetEase Cloud Music’s technical team. Recruitment information is included at the end.