onBackpressureLatest

onBackpressureLatest is an operator in RxJava’s Flowable lineage that implements a specific backpressure strategy focused on delivering only the most recent value when the downstream consumer is slow. It is used to convert a high-frequency upstream into a stream that prioritizes the latest emission, discarding older ones as needed to keep memory usage bounded.

Behavior and semantics

When upstream emits values faster than the downstream can process, onBackpressureLatest buffers at most a single item, the latest one. If a new value arrives before the downstream has requested and received the buffered item, the previously buffered value is replaced. When the downstream requests, the operator emits the currently buffered value (if any) and then continues to buffer and replace with new latest emissions as they arrive. The effect is that subscribers receive the most recent data available at the time of each request, rather than every intermediate value.

Typical use cases

This operator is suitable for streams where only the latest state matters, such as real-time UI updates, sensor readings, or live metrics that rapidly evolve but where processing or display cannot keep up with every event.

Relation to other backpressure strategies

onBackpressureLatest contrasts with strategies like onBackpressureBuffer (which stores all items), onBackpressureDrop (drops new items when full), or onBackpressureError (signals an error). It provides a lightweight way to bound backpressure by prioritizing recency over completeness.

Implementation notes

onBackpressureLatest is provided on Flowable and is part of RxJava’s backpressure toolkit, alongside other strategies such as BUFFER, DROP, LATEST, and ERROR. Developers choose it when the latest value is the only meaningful information to propagate under backpressure.