Event Deduplication

Event deduplication is keeping only one copy of an event when the same event reaches the backend more than once, discarding the repeats.

The basis is a unique identifier attached when the event is created; the backend keeps recently seen identifiers in a window and drops any repeat arrival. Duplicates come from several places: an offline queue retrying after a timeout, the browser restoring a page from its back-forward cache, a user double-clicking a button, and the same action being sent from both the browser and the backend. The mechanism only holds while the identifier is stable, so a retry that generates a fresh identifier defeats it. The real damage from duplicates is silent inflation of conversion counts, because a purchase counted twice looks exactly like growth on a chart.