If you are referring to load balancing HTTP traffic, then we don't support that yet but are actively working on it. This is definitely a must so sit tight!
I have not used those before but it depends on what you want to scale. If it's CPU/memory, you can easily do that. If you need to scale on something else then we should talk! Feel free to open a discussion on this : github.com/kedacore/keda/discussions
Hi guys, just one question for everyone. Let's say we are using Autoscale. After an aks cluster automatically increases the number of nodes, how long does it take for the newly created node to be deleted and downscale, back to the original number (to one node). Of course we are assuming that there is no need (high traffic) for another node to be there. Thank you!
Nice slide about Autiscaling Responsibilities and Control, but why don't you mention that AF with Consumption Plan include cold starts for the newly created instances? I am not sure everyone realizes that only AF Premium Plan iclude the possibility to warm up new instances before they start receving requests, sth AKS with xxx probes can do out of the box ... And then go compare pricing AF Premium : AKS = 2:1 ...
I think that discussion is out of scope here and I'm not the correct person to discuss this, but it's not just the pricing. Although Azure Functions Premium might be a bit more expensive, I don't have to worry about anything other than my application while with AKS you need to ensure your cluster has enough capacity, it's always up, has DR, etc etc. If you are already doing that, then that's a fair choice! If you don't have it yet, go with Azure Functions Premium!
What about combining multiple triggers? If you have two subscriptions, each with its own trigger on messagecount, will they work together? If the first subscription trigger says "not much to do, run one instance only" and the other says "too many messages on my subscription, scale out!", which will win that fight?
@@tomKerkhove That is exactly what I was looking for! And the behavior makes total sense when you think about it. Just got keda2.1 installed in the cluster and can't wait to try it out. Thank you very much!
Great video. Thanks. I have a question like i want to deploy azure functions into AKS using kEDa scaled objects . Problm is i have 4 triggers types in one function , so how can I write the scaled object for the trigger type service bus, event hub, http, timer. Plz guide me how to write
Hi there! KEDA allows you to add one or more triggers. So you can easily add as many triggers that you want, you can learn more here: msft.it/60585hpJm
What's the benefit of running Azure Functions on a Kube Cluster? Is it just portability or is there monetary benefits involved as well. I mean both can auto scale so why should we run Azure Functions on Kubernetes cluster?
This is not a question I can simply answer because it is related to so many topics but one of the main drivers can be companies standardizing on Kubernetes rather than the cloud provider itself. Those function apps will be for free indeed, but you are in charge of keeping them up and running, scaling them, providing capacity, etc while Azure Functions handles all of this for you as a FaaS. Another reason why you could run them on kubernetes is if you run on-prem, edge or multiple clouds where you want to use the Azure Functions paradigm but run it anywhere.. Does that paint the picture a bit for you?
@@tomKerkhove Thank you for responding. It definitely helps. So it's more to do with standardisation of the infrastructure and allow support for hybrid cloud environments right? I mean if we are to stick with Azure only then function apps and azure container apps will work/scale well right?