title: 'useRateLimit' description: 'A hook to do request rate limiting' group: 'Hooks'

An Exobase hook that will check if an incoming request has exceeded the configured maximum number of requests within the configured window. This hook requires a store to be passed and meet the IRateLimitStore interface (having inc and reset functions). You decide how to implement the storage mechanism as long as it surfaces the interface.

Install

yarn add @exobase/use-rate-limit
# or
yarn add @exobase/hooks

Import

import { useCachedResponse } from '@exobase/use-rate-limit'
// or
import { useCachedResponse } from '@exobase/hooks'

Usage

import { compose } from 'radash'
import type { Props } from '@exobase/core'
import { useLambda } from '@exobase/use-lambda'
import { useRateLimit, useServices } from '@exobase/hooks'

type Args = {
  id: string
}

export const findLibraryBookById = async ({ args }: Props<Args>) => {
  return db.libraries.find(args.id)
}

export default compose(
  useNext(),
  useServices({
    store // an objects matching the IRateLimitStore interface (having inc and reset functions)
  }),
  useRateLimit({
    key: 'library.book.by-id',
    limit: {
      window: '5 minutes'
      max: 200,
    }
    toIdentity: (props) => props.request.ip
  }),
  listLibraries
)

Using with Redis store

Create a store module. Here I'm using redis, you can using anything by implementing your own inc and reset functions.

// file: redis-store.ts
import { createClient } from 'redis'

const redis = createClient({
  url: config.redisUrl,
  username: config.redisUsername,
  password: config.redisPassword
})

redis.connect()

// NOTE: This is not the most robust redis
// implementation but makes for a simple
// example.
export const store = {
  inc: async (key: string) => {
    const now = Date.now()
    const created = await redis.setnx(`${key}:start`, `${now}`)
    await redis.setnx(`${key}:count`, 1)
    if (created)
      return {
        timestamp: now,
        count: 1
      }
    return {
      timestamp: await redis.get(`${key}:start`),
      count: await redis.incr(`${key}:count`)
    }
  },
  reset: async (key: string) => {
    await redis.del([`${key}:start`, `${key}:count`])
  }
}

Using with Composition

You probably don't want to import the store in every endpoint function, you can compose it and create your own hook instead.

// file: useRateLimit.ts
import { useRateLimit } from '@exobase/hooks'
import { store } from './redis-store'

/**
 * All functions using this hook will use the
 * configured store and cap requests to a max
 * of 200 per 5 minutes.
 */
export const useRedisRateLimit = (key: string) =>
  useRateLimit({
    key,
    limit: {
      window: '5 minutes'
      max: 200,
    },
    store,
    toIdentity: (props) => props.request.ip
  })

Custom Identity using Authentication

The above examples all use an identity function that references the ip address. In many cases you'll want to identify a requester by their authentication. The toIdentity function has access to the full props object so you can use any request input to create an identity.

import { useRateLimit } from '@exobase/hooks'
import type { TokenAuth } from '@exobase/hooks'
import type { Props } from '@exobase/core'

export const useRateLimitByToken = (key: string) =>
  useRateLimit({
    key,
    limit: {
      window: '24 hours'
      max: 1000,
    },
    toIdentity: (props: Props<{}, {}, TokenAuth>) => props.auth.token.sub
  })

Using with an in-memory store

Create a store object that meets the IRateLimitStore interface and uses a simple variable in memory to track usage.

import { useRateLimit } from '@exobase/hooks'

const db = {}

export const store = {
  inc: async (key: string, timestamp: number) => {
    db[key] = db[key] ?? {
      timestamp,
      count: 0
    }
    db[key].count++
    return db[key]
  },
  reset: async (key: string) => {
    delete db[key]
  }
}

export const useInMemoryRateLimit = (key) =>
  useRateLimit({
    key,
    limit: {
      window: '5 minutes'
      max: 200,
    },
    store,
    toIdentity: (props) => props.request.ip
  })

Using with plan specific limits

Many use cases have specific rate limit rules for a users plan. For example, a free user may be limited to 100 requests/hour while a paid user is limited to 10,000 requests/hour. To support this, you can pass limit as an async function that receives the props as an argument. Use the function form to lookup the limits for the current user.

useRateLimit({
  key,
  limit: async (props: Props<{}, {}, TokenAuth>) => {
    const userId = props.auth.token.sub
    const plan = await db.plans.findForUser(userId)
    return {
      window: '1 hour'
      max: plan.type === 'paid' ? 10_000 : 1_000,
    }
  },
  store,
  toIdentity: (props) => props.request.ip
})