Open jbottum opened 1 year ago
Thanks for the issue Josh! Small question, as I'm now catching up, what exactly do we mean with performance guidelines? Do we have some examples from other projects so that we can start with those and see how to tailor them for Kubeflow?
I think performance in our case is primarily referring to "stress and scale testing".
In a similar vein, Kubernetes has this page that describes "considerations for large clusters" and sets some bounds on what the project is aiming to support.
As per the last discussion, we need to get clarity on "Performance definition" from the CNCF team
Per the CNCF transition, we have been asked to consider developing performance guidelines for Kubeflow. This is a tracking issue for it.