Hi, is there a library I can use for handling rate limiting similar to bottleneck in Node or limiter in Python?

Not that I am aware of. What exactly would it need to do? It shouldn’t be particularly hard to create such a beast. A quick look at these libraries suggest they limit the number of outstanding async requests, right? SWI-Prolog doesn’t have async behavior in this sense. It has preemptively…

Thanks @jan , I need to simulate the following properties from the bottleneck or limiter libs. Basically I need to ensure I am making a max of 100 requests/minute whether single or multi threaded. Any further guidance would be very much appreciated! bottleneck const limiter = new Bottleneck({ "r…

I have no experience with these APIs. They sound a bit similar to the thread pool library, but using an different algorithm to decide when a new task may be started. The thread pool library mostly limits concurrency to avoid overloading a server. Here, it seems you can have some global variabl…

Thank you, this is very helpful. Can you link any guides or examples on implementing the message queue solution? Maybe I can implement it myself… How difficult would this be?

My advice is to use a front-end reverse proxy server that can do both rate limiting and load balancing (and handle DOS attacks, etc.) and (if it makes sense) caching. You don’t want this kind of stuff complicating your server. Rate limiting can’t be easily done with thread pool limits … for example…

Thanks @peter.ludemann but just to clarify, this would be client-side rate limiting. I’m running api calls and hitting the rate limiting on the target endpoint so what I’m trying to do is rate limit my client to a certain number of max requests per minute, not defensively rate limit requests to my o…

[image] m_ac: client-side rate limiting My apologies in misunderstanding what you were asking – I didn’t follow your links. I’m wondering … what does your backend do if you exceed the rate limit? If it returns some kind of “rate limit exceeded” message, there are classic exponential back-o…

No worries. It’s not about what my backend or client do; it’s that the api endpoint I’m hitting will give me 429s (http error for “too many requests”) if I exceed their rate limit. Yes, sleep/1 and get_time/1 . This is what I had in mind. It doesn’t have to be anything nearly as fancy as the nod…

Something like this? (with different intervals, of course): sleeper :- get_time(T0), sleeper(#{last_time: T0, interval:5.0}). sleeper(Time0) :- get_time(Tnow0), stamp_date_time(Tnow0, DateTimeNow0, 'UTC'), writeln(hello:Time0.interval:DateTimeNow0), get_time(Tnow), Slee…

Honestly, I think even this is more complicated than I need. Think I am just going to take the diff of get_time(Time) and get_time(Time0) and then if Time_diff < 60 (or maybe less for padding) we can continue, else sleep. But this has been extremely helpful, thank you both!!

Rate limiting library?

General

brebs April 15, 2024, 3:44pm 13

Global variables and threads points at SWI-Prolog -- Flags

Topic		Replies	Views
More thread wrangling General	4	122	August 6, 2025
Help getting up and running with multithreading General	19	432	May 5, 2024
How to extend maximum of main thread count over 1G? Help!	5	278	January 21, 2024
Options to create a simple side by side asynchronous General	2	260	September 14, 2023
A couple of message queue questions General	3	128	April 28, 2024

Rate limiting library?

Related topics