Design a rate limiter for an external API with a limit of 100 requests per second/minute

To implement a rate limiter with a limit of 100 requests per second or minute on an external API in Go, you can use the Token Bucket or Leaky Bucket algorithm. One of the simplest ways is to use the golang.org/x/time/rate library.

Example with a limit of 100 requests per second:

package main

import (
    "context"
    "fmt"
    "golang.org/x/time/rate"
    "time"
)

func main() {
    limiter := rate.NewLimiter(100, 100) // 100 requests per second, with a "bucket" of 100 tokens

    for i := 0; i < 105; i++ {
        if err := limiter.Wait(context.Background()); err != nil {
            fmt.Println("Wait error:", err)
            return
        }
        fmt.Printf("Request %d sent at %v\n", i+1, time.Now())
    }
}

For a per-minute limit, you can create a limiter with a frequency of 100/60 requests per second:

limiter := rate.NewLimiter(rate.Every(time.Minute/100), 100)

Main points:

• rate.NewLimiter takes the rate and the bucket size.
• The Wait method blocks execution until it is possible to send a request.
• You can use the non-blocking Allow method to check if a request can be sent without waiting.

This approach allows flexible control over the request rate to an external API.