NetTools

API Request Rate Calculator

Convert a raw request count and time window into a per-second rate for capacity planning.

Total requests

Time window (s)

Requests per second

250

Useful for sizing rate limits, load balancers, or checking if you're approaching an API quota.

What you entered

Total requests: 15000Time window (s): 60

1
RPS = total requests ÷ time window
15000 ÷ 60 s= 250 req/s

Result

Requests per second: 250

15000 requests over 60s is 250 requests/second.

Requests per second from a total over a window

Divide total requests by the window in seconds. Rate limits are usually quoted per second, while usage reports come as daily or hourly totals.

Total requests	Window	Requests/second
1,000	1 minute	16.67 rps
5,000	1 second	5,000.00 rps
100,000	1 hour	27.78 rps
1,000,000	1 day	11.57 rps
10,000,000	1 day	115.74 rps

This is the average rate. Real traffic is peaky, so size rate limits and capacity against your busiest minute, often 3-10 times the daily average.

Useful for sizing rate limits

Knowing your actual or expected requests-per-second helps size API rate limits, load balancer capacity, and database connection pools appropriately — under-provisioning causes throttling or errors under real load.

Average rate hides burstiness

A steady 10 req/s average can still include bursts of 100 req/s for a few seconds — for capacity planning, it's often safer to look at peak rate over a short window rather than the average over a long one.

Frequently asked questions

Our API handled 500,000 requests in 24 hours — what's that in requests per second?

About 5.8 requests per second on average. But if 80% of that traffic happens during 8 business hours, peak-hour rate is closer to 17 req/s — plan capacity for the peak, not the daily average.

How do I use this for rate limit planning?

Convert your expected request volume and time window into a per-second rate, then set your rate limit above that with headroom. If normal traffic is 20 req/s, a 50 req/s limit gives healthy burst room without letting runaway clients overwhelm your backend.

What's a normal API request rate?

It varies enormously by use case. A small internal API might see 1-10 req/s. A popular consumer app backend can handle 1,000-10,000+ req/s. Major platforms like Stripe or Twilio serve millions of req/s across their infrastructure.

How does this relate to the uptime/downtime calculator?

This tells you how busy your API is; the uptime calculator tells you how much downtime your SLA allows. Together, they help capacity-plan: a high request rate with a 99.99% SLA means you need redundancy that can handle the load even during partial failures.