Automating In-App Event Schema Validation in Pull Request CI Pipelines

Why Manual Telemetry QA Fails

Relying on manual QA testers to inspect network proxy logs for dozens of event triggers is slow and error-prone. During fast-paced sprint wrap-ups, tracking verification is routinely skipped in favor of core UI testing.

The result is that tracking bugs slip into production releases, where they corrupt telemetry data until the next multi-week app store release cycle.

The Automated Schema Contract Pipeline

Vision Spire Hub implements automated telemetry testing using a three-tier gate:

Tier 1: Static Code Linting (Pre-Commit & PR Check)

AST (Abstract Syntax Tree) parsers scan Swift, Kotlin, and TypeScript code to verify that all invoked event names match keys in the schema catalog.
Disallows hardcoded string literals for event names and property keys.

Tier 2: Unit Test Mock Verifiers

Unit tests trigger synthetic user actions and intercept dispatched event payloads via a local mock logger.
The mock logger validates payloads against the central JSON Schema definitions, failing the test if required properties are omitted or have incorrect data types.

Tier 3: Network Interception Smoke Tests

Automated UI tests (XCUITest / Espresso / Maestro) run key user journeys while capturing HTTP requests through an embedded proxy.
Payloads sent to destination analytics endpoints are matched against baseline gold files.