Hello! I was looking at what are the best practices for running Dagster in K8s with the K8sRunLauncher at scale, I am looking to benchmark Dagster against Kubeflow, running 10, 20,50,100 concurrent workflows

Hi Javier, the two main limiting factors that we’d expect are • The size of your cluster. Obviously your cluster will need to be able to handle 100 pods. It’s also good to take a look at resource requests and limits as a way to help the scheduler distribute the pods among nodes. • The database. At scale we generally don’t recommend using the in cluster postres, it’s more provided for easy spin up. For our deployments we use the managed AWS and GCP options.