Maborak product engineering · ASIN Watcher
Collecting marketplace prices and delivering alerts
The engineering challenge
Price monitoring combines recurring collection, regional price differences and timely notifications. A useful comparison must account for shipping and import fees, while collection workloads must stay separate from the customer application.
Separate collection from customer requests
Worker separation gives collection its own resource budget, but introduces queue coordination and a delay between a marketplace change and its next observation.
Platform architecture
Application, data and infrastructure.
Customer application, collection services, backend APIs, price history, billing and deployment configuration.
- CollectionQueue workers and browser service
- ApplicationProduct tracking and alert rules
- DataPrice history and regional costs
- AccountsBilling and notifications
Data flow
- Product tracking
- Collection workers
- Price history
- Alert evaluation
- Notifications
Collection infrastructure
External workers fetch queue tasks and submit results through an authenticated API. A separate browser service reuses browser processes, isolates request contexts and limits concurrent work.
Application and data
The FastAPI backend manages products, regional pricing, historical observations and alert rules. SQLAlchemy persistence supports separate read and write engines; the deployment stack includes PostgreSQL and connection pooling.
Commercial operations
Accounts, credits, payment integrations and recipient verification sit alongside the monitoring engine. Price events trigger alert evaluation and delivery through configured notification channels.
- FastAPI
- React / TypeScript
- PostgreSQL
- Browser workers
- Container deployment
The resulting system
An integrated monitoring platform that carries a product from scheduled collection through stored price history to a customer alert.
Scope: Customer application, collection services, backend APIs, price history, billing and deployment configuration.
