Nothing to do with this video, all of them you make are extraordinary indeed, but just wanted to say, you are killing us, when I say us I mean us (Java Developers). You really sometimes make me believe that nothing can be compared with C#... Yes, yes... I know the old good saying "You know, everything has its own purpose, and depends where and how you use it. :)", Again, Brilliant work Nick. Euxaristume!
Nick, you are amazing, is incredible how you can be serious and funny at the moment of teach. And that gives a lot of points, because when i see your videos you have all my atention. Keep coding ¡Eres el mejor! :)
I keep telling my self where did I see you, your voice and actions is very familiar, today I found it. you are pretty much the same as Ivar the boneless in the Viking series. thanks for the video I appreciate the help
Something I would like to see is how to consume a rate limited API. There's tons of stuff on creating rate limited APIs for others to consume, not much from the consumer's side.
Really liked your explanation. Just 2 questions. Can we use MongoDB instead of Redis? How can we see the number of requests per client for reporting or monitoring purposes?
First of all, thank you so much for your videos! They are very helpfull. 👏👏 So, I'm trying use this package and I follow everything as you show but when I make a request for my API I didn't get the response headers related to the throttling. Any idea what could be? I missing something? It as a simple ASP Web API .NET 5 project. Thank you!
Always love your content. I have a question here. Can i do such configuration at IIS level, like defining route based rate limiting based on client ip. I ended up creating an action filter that interacts with redis to achieve distributed ip based rate limiting and using that filter on action methods.
It's a middleware that does a lookup every time a request is coming in, so yeah it will but there is no way this feature wouldn't impact performance so it's necessary evil.
Since the header can only be set by the load balancer any user input would be overridden so it would be impossible to workaround if configured properly. It’s equally easy to spoof your request IP as a workaround
That's a solid point. I should have mentioned that rate limiting can also happen on the load balancer which means that the app's performance doesn't have to take a hit. Thanks for raising it
@@nickchapsas would be curious to know the demoed redis backed rate-limit libraries scale limitations are for distributed systems. I also use Apim at work, but it can barely keep up with the throughput our customers demand. A couple families could live quite comfortably on what we spend on apim a month as is, so alternatives are always interesting.
@@ClAddict Redis can scale alongside your application's needs with relativelyminimal medal behind it. I would personally go with a managed Redis cluster and let it rip
Good stuff. But how can a consumer now rate limit itself to this rate limited API? Across multiple nodes? Know of a nuget package that is this easy to wire into httpclient? Because we’ve rolled our own. Badly.
Yeah, I think Nick is right. You can inspect the HTTP response status code (429 Too many requests) and the headers returned from the API. Maybe you can utilize Polly and it’s retry policy to attempt to re-send the http request on the client side of things.
No for 2 reasons. Firstly, .NET supports runtime update of appsettings so depending on how you manage settings you don't need to redeploy. Secondly, the package supports runtime changes on all those settings so you could save them in a db and load them on startup and update them while the app is live
