What I don't understand is since it's the same service can't we just run a performance test, determine the ballpark of the needed resources and just have them capped like in k8s? why would some servers be more powerful than others in the first place?
One perfectly homogenous server gets 20MB POST with image that then needs to be reencoded/cropped
Other gets tiny GET for CSS.
Boom, now your equal servers are inequally loaded. The beauty of leastconn is that if 1st server is loaded it will just naturally get less connections than the lucky ones that only got easy jobs
Yeah, this for sure. Traffic mix changes over the life of an application. In k8s performance isolation certainly isn't a solved problem, and noisy neighbours can affect the "power" of a server.
Any change to the application at all could upset the delicate balance of its performance. You would need to be doing this sort of profiling on every change, using up to date and accurate production traffic.
It ends up being much easier to put some smarts in your load balancer.
Fun fact: Yerba Mate is widely used in Syria and Lebanon due to trade between Latin America and the Levant. I grew up drinking Yerba Mate Tea almost every day.
Yes. It used to be that a great place to leverage this fun fact to your advantage if you were looking to score yerba in the U.S. was a middle-eastern grocery. Also fun to have a bag of Cruz de Malta with an ingredients label in Arabic slapped on it. Nowadays, of course, it's easier just to go online.
