Skip to content

Flow operators and backpressure

← Back to Concurrency

Flow Operators and Backpressure Deep Dive

Overview

Operator selection controls cancellation semantics, buffering behavior, and user-visible responsiveness in stream pipelines.

Core Concepts

  • collectLatest for latest-only rendering
  • flatMapLatest for source switching
  • buffer and conflate for producer/consumer decoupling
  • backpressure policy as product behavior decision

Internal Implementation

Operators insert coroutine boundaries and channels in the pipeline. collectLatest cancels previous collector block on new emissions. flatMapLatest cancels previous inner flow when outer source updates. buffer introduces queueing; conflate keeps only newest queued value.

Threading Model

Operator chains may hop dispatchers (flowOn) and create additional scheduling costs. Keep context shifts explicit to avoid hidden overhead.

Coroutine / Flow Behavior

Latest operators improve UI freshness but can drop intermediate work. Buffered chains improve throughput but may increase latency under bursts.

Code Examples

queryFlow
    .debounce(250)
    .flatMapLatest { query -> repository.search(query) }
    .flowOn(Dispatchers.Default)
    .collectLatest { result ->
        render(result)
    }

Common Interview Questions

  • Q: Why does collectLatest cancel in-flight work? A: Start from delivery semantics: use StateFlow for durable state, SharedFlow or Channel for transient events, and lifecycle-aware collection to prevent duplicate work.
  • Q: When is flatMapMerge better than flatMapLatest? A: Answer with correctness first and throughput second: cancellation model, dispatcher choice, bounded parallelism, and contention or latency measurements.
  • Q: What tradeoff does conflate make? A: Answer with correctness first and throughput second: cancellation model, dispatcher choice, bounded parallelism, and contention or latency measurements.
  • Q: How do you debug dropped intermediate emissions? A: State load and SLO assumptions first, identify the first bottleneck, choose scaling and consistency strategy, and explain fallback behavior for partial failures.

Production Considerations

  • choose operator semantics per UX expectation
  • cap buffers for bursty sources
  • instrument cancellation and processing latency
  • avoid accidental heavy work on main dispatcher

Performance Insights

Pipeline performance is dominated by work per emission and queue behavior, not operator count alone.

Senior-Level Insights

Strong answers connect operator semantics to explicit product tradeoffs: freshness, completeness, and cost.