Hey my team is migrating our Metaflow setup from A...
# ask-metaflow
a
Hey my team is migrating our Metaflow setup from AWS Batch -> AWS EKS. My company already uses AWS EKS heavily and my team recently deployed Jupyter Hub on said cluster. But noticed we are getting
InsufficientInstanceCapacity
when trying to provision GPU instances(g5.xlarge), even though the cluster is deployed in us-east-1 with multi-AZ’s. Anyone else facing issues getting sufficient GPUs for your workflows? I do not face the same issue when requesting the same instances on AWS Batch/Sagemaker Endpoints.