Migrating AWS EC2 auto scaling group for RabbitMQ cluster with minimal downtime

[amin testing]

Hello,

I have an AWS EC2 auto scaling group that creates 3 EC2 instances in my cluster. I am using these instances to run a helm chart of my RabbitMQ cluster. (rabbitmq-0, rabbitmq-1, rabbitmq-2).

My issue is that the instance type of my auto scaling group did not satisfy my environment as I need more specs such as increase storage size of the nodes and etc. I am confused between two quick fixes for updating my specs with minimal downtime for my RabbitMQ cluster. Either I change the auto scaling group which will create new node instances that my rabbitmq pods will join after the old nodes get terminated or should I create a new auto scaling group with new node instances (with the updated specs) then change the Statefulset node-selector of my rabbitmq to join the new node instances. Which solution do you recommend that can give me lower downtime for my RabbitMQ cluster?

Please advice me regarding this scenario as your feedback is very important for me and highly appreciated.