What is Rate Limiting / API Throttling? | System Design Concepts

API Throttling or Rate Limiting is a concept used to describe how applications control the amount of traffic their APIs support. Rate Limiting works but the server identifying clients, and applying an algorithm that keeps that caller's traffic below a certain threshold (i.e. 5 calls per second). In this video, I talk about what api throttling / rate limiting are, why applications rate limit, techniques, and implications.

📚 MY RECOMMENDED READING LIST FOR SOFTWARE DEVELOPERS📚

🎙 MY RECORDING EQUIPMENT 🎙

💻 MY DESKTOP EQUIPMENT 💻

🌎 Find me here:

#SoftwareDevelopment
#Throttling
#RateLimit4

[HNVIDEOS]

Such an intelligent Engineer you are bro, best explanation ever.

[kugurerdem]

The client perspective was especially helpful in my case. Thanks!

[williamchan]

very clear explanation! Recently I need to dive into the details of API during my daily work. It really helps!

[Hrit]

Thanks for the clear explanation. Very beginner friendly.

[bernhardriegler]

Excellent explanation and hood pointers to existing libraries. Thank you!

[rajeshrenke]

7:56 it says bucket is empty and request coming in this time is going to be rejected. I didn’t quite catch that. If the bucket is empty it means it can accommodate 100 request, which is the size of the bucket, and should be served. Isn’t it. I want to know what am I missing?

[jinny]

Your explanation is so clear to understand than ever. Please keep uploading system design videos. I just subscribed to your channel 👀

[sulaymanmuhammedjawneh]

Great delivery. I enjoyed and benefitted from it

[shahamitsynygy]

Nice video. You have explained the concept very well. One question - Isn't this already handled by API management softwares/solutions along with other api management features?

Also any plans to create a video on api management aspects?

[spyderwon]

So i'm not clear what the different between rate limiting and throttling actually is. Are they interchangeable?

[malikbrahimi]

How do you deal with 3rd party rate limits in your own application? Do you have to just create a distributed system to keep retrying against the 3rd party API?

[jnsmith]

I'm hitting a scenario where Azure is talking on the outbound too fast(too many times in a minute). As a result Azure's token will eventually fail\go into Quarantine mode. I see that I can limit rates on the inbound per most examples. So, I'd view this as limiting part of the roundtrip communication.

Question- Does rate limiting allow me to stretch out the overall communication process to a longer period of time ? I'm syncing users to a Cloud based app. I'm trying to limit the updating\creating process to under 400 requests per minute.

So, would applying the rate limiting on the inbound only slow down the overall process ?
Is there a way to apply rate limiting on the outbound direction ?

[gfboxp-]

I try to sign in coinbase pro this message shows what is this mean?? Wait tilland retry?? Please help.."Rate limited have been exceeded, please try again in the moment.This error occurs when you exceeded number of allowed requests our backend over a certain time period.

[MauriceWilliams]

This was pretty awesome. I'm guessing that you know how to code as well?