Skip to content

Professional engineering work Confidential · source not public

Event-Driven Data Processing

High-volume, queue-based asynchronous processing of product, stock and price data.

01 · Problem

Product, stock and price updates from sellers arrive in sudden bursts. Handling them synchronously slows the API and couples systems together.

02 · Approach

Turn updates into messages distributed to independent consumers over RabbitMQ, write every consumer to tolerate reprocessing, and keep the search index and cache current through events.

03 · Technical challenges

  • Protecting the rest of the system during bursts
  • Consistency when a message is processed more than once
  • Keeping database, cache and search index in sync

04 · Value delivered

A scalable, more predictable data flow where heavy processing is decoupled from user-facing APIs.