Testing Professor
Testing ProfessorMulti-Cloud Guidebook · AWS
All scenarios

End-to-End QA Scenario

Mobile Automation Across 30 Real Devices

CodeBuild builds signed IPA/APK; Device Farm executes Appium suite across a curated device pool; artifacts archived to S3.

Architecture

CodePipeline ──► CodeBuild (build IPA/APK)
                       │
                       ├─► S3 (signed artifacts)
                       │
                       └─► CodeBuild (kick off Device Farm run)
                                  │
                                  ▼
                           AWS Device Farm (30 devices, Appium suite)
                                  │
                                  ├─► S3 (videos, logs, screenshots)
                                  └─► SNS ─► Slack (#qa-mobile)

Workflow steps

  1. 1

    Secrets

    Signing keys + App Store credentials stored in Secrets Manager; CodeBuild reads at runtime.

  2. 2

    Build

    CodeBuild produces signed IPA + APK and uploads to S3 with version metadata.

  3. 3

    Automation matrix

    Second CodeBuild stage schedules a Device Farm run across a curated device pool with parallel execution.

  4. 4

    Collect

    Artifacts (video, syslog, perf) mirrored to S3 beyond Device Farm's default 30-day retention.

  5. 5

    Report

    Lambda polls run status, summarizes pass/fail per device, posts to Slack via SNS.

Key takeaways

  • Real-device automation catches OEM-specific bugs emulators miss.
  • Mirror artifacts to S3 — provider retention is shorter than your audit need.
  • Per-device durations expose perf regressions before users complain.