Optimizing Gray Release Process and Automation at DeWu App

DeWu’s release cycle originally followed a fixed bi‑weekly iteration, with occasional single‑week hot‑fix insertions. The long lead time from gray‑release preparation to market launch caused delays in business feedback, A/B testing, and feature validation.

To improve efficiency, the team first quantified release data (real‑time monitoring, effectiveness evaluation, problem diagnosis, continuous optimization). Metrics such as gray‑release coverage, crash‑affected users, and time‑to‑market were collected in version and quarterly reports.

Based on the analysis, several optimization directions were defined:

Standardize release rhythm and shift service‑side publishing and client packaging earlier by four hours.

Introduce automated capabilities in the workflow platform to reduce manual effort and errors.

Key automation features :

Automatic creation of release plans with inferred version parameters.

Pre‑release checkpoints (privacy compliance, code merge, crash fixes, package size, third‑party approvals).

Dependency reminders for component integration and offline crashes.

Auto‑packaging triggered by dependency satisfaction, scheduled times, or KOL channel requirements.

Test‑regression reminders that push pending QA tasks to Feishu groups.

Event subscription system that publishes important events (build success, plan creation, test start) via Feishu, DMQ, or HTTP.

Automated market‑monitoring for Android stores and a custom iOS App Store Connect workflow.

These improvements shortened overall release delay by more than 70 % and allowed each full‑release version to reach users on average one day earlier. The quarterly and version reports now provide clear visibility into quality, efficiency, and team performance across all functional domains.