Three Postgres sources need CDC

A medium Pipeline Design interview practice problem on DataDriven. Write and execute real pipeline design code with instant grading.

Domain

Pipeline Design

Difficulty

medium

Problem

Three Postgres sources need CDC. orders has 10M writes per day; doubling that with triggers would force a larger instance. shopping_cart deletes are critical to capture and the existing query-based ingestion misses them. legacy_events runs on a hosted Postgres without WAL access, so log-based CDC is unavailable. The section names three CDC families: trigger-based (captures deletes; doubles writes), query-based (cheap; misses hard deletes), and log-based (low overhead; needs WAL access and slot management). Identify the CDC family by replacing each source's ingest transform with one whose name states the chosen family and the deal-breaker that ruled out the alternatives.

Practice This Problem

Solve this Pipeline Design problem with real code execution. DataDriven runs your solution and grades it automatically.