Throughput Window

Track requests per second over a rolling time window.

rps

Detailed Description

Throughput measures how many units of work your service completes per second. A rolling window smooths short spikes while still reacting fast enough to show traffic changes.

Throughput should always be read together with latency. More RPS is not a win if p95 or p99 latency explodes, because that means the service has crossed its safe capacity point.

Mental Model

record request time
-> remove old samples
-> count/window = RPS

When To Use

->Find the max sustainable RPS
->Compare latency at different load levels
->Detect sudden traffic drops or spikes

Tools: rolling counters, Prometheus counters, autocannon RPS

throughput-window.js

Live Editor

Output

Run code to see output...