High-Concurrency Marketplace (Mov33)

What

A high-traffic event ticketing marketplace.

Who

Event organizers and thousands of concurrent buyers.

Why

Previous solution crashed under load, leading to overselling.

Constraints

  • High concurrency during ticket drops.
  • Zero tolerance for overselling.
  • Real-time seat availability updates.

Architecture

ClientAPI GWAuthCoreAnalyticsStorage

System designed for isolation. Auth, Core, and Analytics scale independently based on load profiles.

Engineering Decisions

Atomic Transactions

Leveraged Firestore transactions to ensure inventory integrity.

Idempotent Webhooks

Designed payment processing webhooks to handle duplicate events gracefully.

Cloud Functions

Offloaded heavy processing (PDF generation, email) to serverless functions.

Frontend & UX Intent

[ UI_SNAPSHOT_PLACEHOLDER ]

Mobile-first checkout flow optimized for speed.

Real-time seat mapOptimistic inventory locking

Outcome & Reflection

Shipped

A stable ticketing platform handling thousands of concurrents.

Improved

Checkout conversion rate increased by stabilizing the backend.

Different

Would consider a SQL database for complex relational queries if defining the schema from scratch again.

    Mark Allan | kihumba.com