2024-01-02
4560
#node
Philip Obosi
14373
Jan 2, 2024 β‹… 16 min read

Understanding and implementing rate limiting in Node.js

Philip Obosi Frontend engineer and data visualist πŸ‘¨πŸ»β€πŸ’» based in Lagos, Nigeria.

Recent posts:

How To Use The CSS Cursor Property

How to use the CSS cursor property

A single line of CSS can change how users feel about your UI. Learn how to leverage the cursor property to signal intent, improve interaction flow, and elevate accessibility.

Chizaram Ken
May 28, 2025 β‹… 6 min read
Build TypeScript App Vite

How to build a React + TypeScript app with Vite

We explore the benefits of building an app with React, TypeScript, and Vite, and compare its performance to the same app built with CRA.

Clara Ekekenta
May 28, 2025 β‹… 7 min read

How to use Claude to build a web app

Learn how to build a weather app using Claude, from setting up infrastructure to creating a functional UI that displays city-based forecasts.

Andrew Evans
May 28, 2025 β‹… 8 min read
Enhancing LLMs With Function Calling And The OpenAI API

Enhancing LLMs with function calling and the OpenAI API

Implement a scheduling assistant using the OpenAI API and function calling, no Model Context Procedure (MCP) required.

Kapeel Kokane
May 28, 2025 β‹… 11 min read
View all posts

11 Replies to "Understanding and implementing rate limiting in Node.js"

  1. 2 of 3 cons of fixed window counter are not fair:
    – “user’s window should start counting from the time of their first request” -> this is easy to implement.
    – “burst traffic towards the end of a window” -> it may be issue, if your service is for one customer. It is unlikely, that all your thousands users would make all requests at once.

  2. Hi,
    It looks like using app.use() would limit the rate to the whole API. How would you go about applying rate limit to only a particular POST request while letting users do unlimited GET requests?

  3. Michal,

    You can do this by applying the middleware to the POST route directly instead of `app.use`

    e.g.

    `app.post(‘/limitedRoute’, customRedisRateLimiter, (req, res, next) => {})`

  4. When the record is null in the Redis store, you create the record, store it and then go to the next middleware. Shouldn’t there be a return statement after the next() instruction to prevent the middleware from executing the rest of the code ?

  5. you should wrap “await redisClient.connect()” in if statement with condition “!redisClient.isReady” or “!redisClient.isOpen” so it doesn’t throw “Socket already opened” error.

  6. this line get time of 24 hours ago from now ‘const windowStartTimestamp = moment().subtract(WINDOW_SIZE_IN_HOURS, ‘hours’).unix();’ and the record in redis already deleted after 24 hours , so how it comes?

  7. I tested the first implementation. I noticed that requestCount is only incremented when you call a different endpoint. But I want the rate to be per request, no matter the endpoints.

Leave a Reply