Is it possible to tag ECS jobs with GPU information?

hey alexander - does that require running in EC2 rather than fargate? Right now our ECS run launcher only supports fargate

right, thanks. Do you know if it can be done on EKS?

I don't offhand, but if it could, you could likely configure the pod to use it using some of the fields here: https://docs.dagster.io/deployment/guides/kubernetes/customizing-your-deployment#job-or-op-kubernetes-configuration

ah, thanks. there doesn't seem to be anything GPU specific on that page but i guess its just a question of tagging the node correctly? EKS does allow it.

yeah, that page is for applying k8s config in general

Do i have to use Helm to do K8s in Dagster?

you don't have to, but it's what we would recommend

is that how the dagster daemon launches job pods?

dagster doesn't use helm internally, no

great 🙂 (I don't like Helm, too obfuscated for my taste)