The question remains: we don't have a time‑based key. Person T1 receives a transaction; Person T2 should immediately know about T1's transaction.

When real‑time visibility is required without a time‑based key, an event‑driven architecture is ideal. As soon as Person T1 processes a transaction, an event is published to a message bus (Kafka, RabbitMQ, or a cloud pub/sub service). Person T2 subscribes to the relevant topic or queue and receives the event instantly. To guarantee immediate delivery, use a durable queue with at‑least‑once semantics and idempotent processing on the consumer side. If strict ordering is needed, partition the topic by a deterministic key (e.g., user ID) so that all events for a user are processed in order. For low latency, consider using a lightweight stream like Redis Streams or a WebSocket push from the event bus. Consistency can be achieved by storing the event in a shared state store (e.g., a distributed cache) that both T1 and T2 can query. This pattern decouples producers and consumers, scales horizontally, and ensures that Person T2 is notified immediately after T1’s transaction is committed.