Multiple replicas of nginx-ingress-controller on AWS EKS

Question

1. Are there any issues to scale replicas:1 to replicas:3 for nginx-ingress-controller deployment?
2. So the controller has created AWS Classic ELB, with many replicas - are we going to have the same (single ELB) or one per POD (as number of PODS will increase)?

The reason I'm asking is that I'm running the controller inside EC2spot-based AWS EKS cluster and when the EC2SpotInstance is interrupted by AWS - I get downtime.

Answer 1

Are there any issues to scale replicas:1 to replicas:3 for nginx-ingress-controller deployment?

Nginx ingress controller watches on few resources exposed by kubernetes API server such as service , endpoints etc. When you scale number of replicas it might put more pressure on the kubernetes API server because now 3 replicas will watch instead of 1.But with increase of 2 replicas it may not be observable but I suggest to put some monitoring in place to observe any impact.

So the controller has created AWS Classic ELB, with many replicas - are we going to have the same (single ELB) or one per POD (as number of PODS will increase)?

Yes you will have same and single ELB because ELB is created per service object. So increase in replica count of pod should not create new ELB. Same ELB will send traffic to the kubernetes service which will perform load balancing to 3 replica pods of nginx ingress controller.