Design a Scalable Catalog Service

# Design a Scalable Catalog Service Design a catalog service that lets authorized producers create and update items while customers browse, retrieve, filter, and search the published catalog. Each item has structured attributes, descriptive text, a publication state, and a version. The service should support frequent reads, less frequent writes, safe schema evolution, and consistent publication of related item data. State any assumed scale and freshness targets. ### Clarifying Questions to Ask - Is this a product catalog, media catalog, internal service catalog, or another domain? - Which fields are strongly consistent, and how stale may browse and search results be? - Are variants, regional availability, prices, inventory, or permissions inside this service's boundary? - What write approval, version history, and rollback behavior do producers need? ### What a Strong Answer Covers - Item and version data models, producer and customer APIs, validation, and publication state transitions. - A source-of-truth store plus cache, browse index, and search index projections. - Idempotent updates, outbox or change-stream propagation, stale-read behavior, and reindexing. - Partitioning, hot-item handling, access control, observability, failure recovery, and schema migration. ### Follow-up Questions 1. How would you publish a coordinated update to thousands of related items without showing a half-updated collection? 2. What happens when the search index is unavailable while catalog writes continue? 3. How would you introduce a new required attribute without breaking older item versions?