Booking System QA & Automated Test Suite by Yusuke MiyajimaBooking System QA & Automated Test Suite by Yusuke Miyajima

Booking System QA & Automated Test Suite

Portfolio project
Yusuke Miyajima
Business outcome
A booking team can release seat holds, confirmations, payments, cancellations, refunds, capacity changes, and session rules with reproducible evidence against overselling, duplicate charges, broken authorization, stale holds, and concurrency failures.
Evidence
107 of 107 API and browser checks passed with retries disabled; all 6 seeded defects were detected; and two real defects were reproduced, fixed, and pinned with regression coverage.
Execution evidence: 107 clean checks with retries disabled, six controlled defect detections, and a reproducible P1 keep-alive defect report fixed with regression coverage.
Execution evidence: 107 clean checks with retries disabled, six controlled defect detections, and a reproducible P1 keep-alive defect report fixed with regression coverage.
The assignment
I treated this as a complete QA engagement rather than only a Playwright repository. The controlled Seatly booking application covers workshop browsing, seat holds, confirmation, simulated payment, cancellation, refunds, administrator capacity, and session rules.
Risk-based approach
I mapped the highest-impact failures first: overselling, broken authorization, duplicate charging, stale holds, refund integrity, malformed requests, and capacity races. The deliverables include a risk assessment, test strategy, 12 structured manual cases, API and browser automation, defect reports, and an execution report.
Coverage and evidence
• 101 API regression checks • 6 Chromium end-to-end journeys • 107 of 107 clean checks passed • 0 retries, skipped tests, or flaky results • 6 of 6 controlled seeded defects detected
Defects found and fixed
The engagement found two real defects. A blank administrator capacity was converted to zero and silently closed a workshop. An oversized HTTP request left unread bytes on a reused socket, causing a later request to hang. Both defects were fixed in commit c7345f9 and pinned with regression coverage.
Why the evidence is trustworthy
The second defect was reproduced with a plain Node keep-alive client independently of Playwright. Controlled defect mutations prove the suite fails when key protections are broken instead of merely producing a green report.
Handoff
The repository includes the application under test, test strategy, risk assessment, manual cases, automation, reproducible bug reports, CI configuration, and operator commands. Known limitations are documented explicitly.
Completed work
54-second walkthrough: real booking flow, confirmed payment state, and reproducible QA evidence.
Need reproducible release evidence? My Risk-Based QA, Regression & Release Readiness service covers risk analysis, test design, automation, defect reproduction, and release recommendations.

0 likes · 1 view · Aug 26, 2026

Built a practical QA suite covering UI workflows, API behavior, regression risks, and reproducible test reporting.

Tags