AWS EKS Managed Node groups to Fargate

Hi

I am currently running my workload on EKS using Managed node groups. I am looking at options to leverage Fagate for my EKS cluster. I currently run both stateless and stateful workloads using NLB. I would like to make this migration from managed to Fargate as a Zero down time migration without impacting my customers.

This is my thinking at a high level,

a. Create a New Fargate profile on my existing 1.20 cluster with required selectors with required namespaces b. Taint the Manged worker nodes and Drain the pods

My questions

a) Has anyone done such migrations? Will you be able to share your experiences / any gotchas? b) I have 100s namespaces, Looks like there is a limitation on the number of selectors you can have on the Fargate profile. Can I specify all the namespaces in a single profile? Any known limitations in creating number of profiles? c) CSI Drive installation seems to be a manual step on the existing clusters. Any experiences?

Thank you