Rancher Cluster Monitoring - Ignore Fargate Nodes

I’m running Rancher on EKS and would like to be able to use Fargate as a compute resource.

However, I’ve noticed that whenever I have pods running on Fargate servers, that the Rancher’s Monitoring app gets stuck trying to deploy exporter-node-cluster-monitoring and grafana-cluster-monitoring. I think this is because Fargate will only allow 1 pod to run on an instance.

Is there any way to either tell Rancher Monitoring to ignore Fargate nodes?

Is there any way to annotate nodes such that they would be ignored for the purposes of monitoring?

Many thanks!


Use this rule in Node Scheduling:

eks.amazonaws.com/compute-type != fargate