Average vs peak load
Size your servers for the peak, not the average, or traffic spikes will overwhelm you.
Home / Dev & AI / API & Backend Cost / API Requests Per Second Calculator
API & Backend Cost
API Requests Per Second Calculator
Find your real throughput. Enter monthly requests and a peak to average ratio, and see the average and peak requests per second you need to design for.
Average requests / second
$0
Peak requests / second—
Requests per day—
Requests per minute (avg)—
Requests per month—
Peak requests / minute—
What the API Requests Per Second Calculator does
Requests per second, or RPS, is the number engineers use to size servers, set rate limits and plan capacity. The catch is that traffic is never flat. Average RPS from your monthly total tells only part of the story, because real usage peaks during busy hours and events, often several times the average.
This calculator converts your monthly request total into an average RPS, then applies a peak to average ratio to estimate the peak you must actually handle. Sizing for the peak, not the average, is what keeps an API responsive when traffic surges.
Heads up on pricing. The rates built into this tool are public list estimates for July 2026 and can change without notice. Providers also offer batch discounts, prompt caching and volume tiers that lower real costs, so treat the output as a planning estimate, not a quote.
The formula
average RPS = monthly requests ÷ 2,592,000
peak RPS = average RPS × peak to average ratio
peak RPS = average RPS × peak to average ratio
Worked example
For 200,000,000 requests a month, average RPS is 200,000,000 ÷ 2,592,000 = about 77 RPS. With a peak to average ratio of 3, you should design for roughly 231 RPS at the busiest moments.
Ways to lower your API and backend costs
- Serve responses through a CDN. Cached, edge-delivered responses cut egress cost and take load off your origin.
- Compress payloads. gzip or brotli plus trimming unused JSON fields reduces bandwidth on every single request.
- Cache common responses. Avoid recomputing identical results for repeated reads, which lowers both compute and request fees.
- Reduce chatty designs. Combine several small calls into one batch endpoint or a single GraphQL query to cut request volume.
- Right size and autoscale. Match capacity to real peak traffic instead of paying for servers that sit idle most of the day.
Frequently asked questions
How do I calculate requests per second?
Divide your monthly requests by the seconds in a month (2,592,000). That gives average RPS. Multiply by a peak ratio to estimate peak RPS.
What is a good peak to average ratio?
Consumer apps often see 2 to 4 times average at peak, while spiky or event driven traffic can hit 10 times or more. Use your own analytics if you have them.
Why size for peak instead of average?
Because your servers must handle the busiest second without falling over. Sizing for the average guarantees outages during normal daily peaks.
How does RPS relate to server count?
Peak RPS combined with response time gives required concurrency, which sets how many servers you need. The server capacity calculator does that step.
Is the monthly conversion exact?
It uses a 30 day month for consistency. Real months vary slightly, but the difference is negligible for capacity planning.