How Qunar Reduced Fault Rate and Boosted Development Efficiency with Automated Testing, Intelligent Recommendation, and Localization

Qunar, a large travel service platform, faced a massive testing workload with over 10,000 checklist items and high release‑fault rates; to address this, they introduced a left‑shift functional testing strategy that replaces manual, repetitive QA tasks with an automated testing platform.

The platform uses code‑coverage probes and dynamic checklist generation to automatically create and maintain test cases, enabling massive parallel execution, environment‑aware deployment, and real‑time result assertion, which reduced regression testing time from half a day to under five minutes and lowered QA scheduling from 45% to 15% of project effort.

To further decrease fault rates, Qunar built an intelligent recommendation system that captures code‑coverage data, filters irrelevant changes, and automatically recommends the minimal set of test cases needed for a given code change, achieving a fault‑rate reduction from 3.8% to below 1% and maintaining an overall release fault rate under 0.4‰.

Finally, a localization platform was created to bring beta‑environment configurations into developers' local environments via soft‑routing and IDEA plugins, allowing developers to run and validate tests locally, increasing self‑testing adoption from 0% to over 70% and cutting project turnaround time by more than a third.

The combined practices delivered a holistic improvement: automated testing, precise fault detection, and localized development together enhanced testing coverage, reduced release failures, and accelerated development cycles, with future plans focusing on cloud‑based development, extensible tooling, and deeper IDE integration.