System Design · Karat · Medium
Whenever a user signs a document, the system attempts to send a notification. The problem is that when one of these sends does not go through, nothing records the failure, so there is currently no log to consult. Design a scalable approach for identifying every document whose notification failed to send. Your design should address three aspects: Backfill: recovering the set of affected documents from the history that already exists. Ongoing detection: catching new failures…