sparse-intern-71089
09/26/2022, 6:44 AMlimited-rainbow-51650
09/26/2022, 6:49 AMrhythmic-whale-48997
09/26/2022, 6:57 AMrhythmic-whale-48997
09/26/2022, 6:58 AMlimited-rainbow-51650
09/26/2022, 7:00 AMrhythmic-whale-48997
09/26/2022, 7:04 AMechoing-dinner-19531
09/26/2022, 7:07 AMlimited-rainbow-51650
09/26/2022, 7:15 AMNodeGroup
which has default instance count set to 0
but is allowed to scale up and down.
• schedule your workload as a K8s Job with special pod placement tactics so it can only end up on the special EC2 NodeGroup
When the job is scheduled, EKS will first scale up with 1 instances to host your job, and after the job is run, you allow the NodeGroup to scale down again.
Here is a link to an AWS workshop explaining this:
https://www.eksworkshop.com/beginner/140_assigning_pods/rhythmic-whale-48997
09/26/2022, 7:17 AM