Little's Law Calculator

How to use

1. 1Pick what you want to solve for: L (concurrency), lambda (throughput), or W (time in system). The locked field on the form shows which side is being computed.
2. 2Fill in the two known sides. Each numeric field has a unit picker so per-second, per-minute, per-hour, per-day throughput and millisecond through day response time all work without manual conversion.
3. 3Read the headline result on the right with its full unit breakdown. Throughput is shown per second, minute, hour, and day; response time is shown in ms, s, min, and hr where relevant.
4. 4Optionally type a service rate mu (the maximum rate of one server or one pool) to see the M/M/1 utilization rho = lambda / mu. The band is color coded so you can tell stable from saturated at a glance.
5. 5Try a preset to reproduce a typical capacity question (web server, queue worker, support team, response time, capacity, checkout queue) and tune from there.
6. 6Click Copy report to grab a plain-text summary with the inputs, the headline, every unit conversion, and the utilization check, ready to paste into a runbook, design doc, or load-test ticket.

About this tool

Little's Law Calculator turns the relationship L = lambda x W into the three capacity questions a backend or operations team actually asks. Pick what you want to solve for. Concurrency L gives the average number of items in the system at once, useful for sizing thread pools, connection pools, queue workers, support agents, and any other resource that has to be present while an item is being handled. Throughput lambda gives the sustainable arrival or completion rate at a given concurrency and response time, useful for setting a load-test target or sanity checking a published SLA. Time in system W gives the average end-to-end response time at a given concurrency and throughput, useful for translating a thread-count number from a dashboard into a latency number a product team understands. Throughput accepts per-second, per-minute, per-hour, and per-day inputs; time in system accepts milliseconds, seconds, minutes, hours, and days; results are reported in every relevant unit so the answer is ready for a runbook, a Slack message, a Jira ticket, or a capacity model spreadsheet. An optional service rate mu unlocks an M/M/1 utilization sanity check: rho = lambda / mu is shown with a green stable band below 70%, an amber caution band from 70% to 85%, and a red high-risk band above 85% (with an unstable warning when arrivals exceed the service rate). Six built-in presets demonstrate the most common shapes: web server thread sizing, queue worker count, support team headcount, solve-for response time, solve-for capacity, and a checkout queue with utilization. Little's Law applies in steady state for any queueing discipline and any arrival or service time distribution, so the result is robust whether the work is a CPU-bound microservice, a slow batch ETL job, or a person on a phone. A copy-friendly plain-text report includes the inputs, the headline result in every unit, and the utilization band so the math is reproducible and reviewable. Useful for SREs sizing thread pools, backend engineers tuning database connection counts, performance engineers planning load tests, support managers staffing the queue, product managers translating latency targets into capacity, and students learning operations research and queueing theory. All math runs in your browser; the numbers you enter never leave your device.

Free to use. Works in your browser. No signup, no login.