Quickstart

Rate limiting is built in: configure token buckets, sliding windows, or adaptive limits that respond to backpressure.

Batching is automatic—control it via `batch(100)` or `batch({ size: 50, timeoutMs: 1000 })` to trade latency for throughput.

const result = await stream.request('POST', '/transform',  data: payload ,  timeout: 5000 );

Metrics are published to standard formats (Prometheus, CloudWatch, DataDog) with zero configuration.

Quickstart

Metrics are published to standard formats (Prometheus, CloudWatch, DataDog) with zero configuration.

stream.pipe(map(e => e.data), buffer(100, 5000), flatMap(batch => fetch('/api/bulk',  body: JSON.stringify(batch) ))).catch(console.error);

Graceful shutdown waits for in-flight events to complete, then closes connections cleanly—no lost events, no crashes.