Additional worker deployments using `KubeCluster`


Similar to the feature request in Deploying Multiple Worker Types with Helm for the dask helm chart, I was wondering whether you’re also thinking about adding the possibility to deploy additional workers (e.g. GPU nodes) with dask-kubernetesKubeCluster to allow for ephemeral/temporal dask cluster deployments or whether you could think of other options of achieving this?


@bstadlbauer Welcome! There’s been some work on this as a part of the Dask Operator for Kubernetes initiative, here: Dask Operator — Dask Kubernetes 2021.03.0+65.g3967206 documentation. :smile:

Ref GH issue: Kubernetes Operator · Issue #256 · dask/dask-kubernetes · GitHub

1 Like