Umeng App Core Metrics and Intelligent Patrol

Problem

Umeng app data is often spread across new users, active users, launches, retention, channel, and version metrics. Investigating anomalies usually requires manually calling several APIs, aligning dates, and tracking the correct APPKEY. umeng-api packages these queries into a reusable Python SDK, making it easier to pull core statistics and anomaly reports from scripts, internal tools, or agent workflows.

How It Works

• Credential loading: It prefers values passed in code, then reads umeng-config.json, then falls back to environment variables, keeping apiKey and apiSecurity out of business logic where possible.
• Metric queries: For a target APPKEY and date, it calls Umeng UApp APIs for new users, active users, launches, usage duration, retention, channel, and version data.
• Anomaly inspection: Beyond baseline metrics, it queries anomaly-detection and attribution reports to help identify likely business causes behind metric shifts.
• Error handling: It distinguishes aop.ApiError, aop.AopError, and ValueError, which helps separate gateway failures, pre-request client errors, and missing credential configuration.

Boundaries

The skill assumes a valid Umeng account, app APPKEY, and API credentials. It does not replace monitoring, alerting, or incident-response systems, and it remains subject to Umeng API rate limits. Sensitive credentials should not be hard-coded, and configuration files should be excluded from version control and file-permission protected.