Update VPC ENI/IP limits table to add r5, r5d, z1d and t3 instance types

[cablespaghetti]

Description of changes:
Add new instances types so users of them don't get messages such as [ERROR] Failed to get eni limit due to unknown instance type r5.large and a failure to assign any interfaces to pods.

[vsiddharth]

@cablespaghetti Thanks a lot for submitting this pull-request. Your changes look good to me.

However, could you please confirm that this contribution is made under the terms of the Apache 2.0 license?

Additionally, could you also describe any tests that you performed to validate the changes??

[cablespaghetti]

@vsiddharth Sure. My contribution is made under the Apache 2.0 license.

I have not been able to get a working build environment unfortunately, so have not been able to test my changes. However I have very carefully checked them against https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-eni.html#AvailableIpPerENI

Thanks

[vsiddharth]

@cablespaghetti Thanks a lot for your contribution.

Please let us know what issues you are facing with the build environment.

[pahud]

Do we need to wait for the aws-vnc-cni-1.2.0 being released until we can use those new instance types?

[doxavore]

I can confirm when taking a broken system using t3.medium instances to t2.medium, this is fixed. It does appear that until a new version is released (and perhaps the aws-ami picks it up) this is still a problem.

[Jeeppler]

I was able to successfully use a t3.large in my EKS cluster.

[msvechla]

@Jeeppler what AMI did you use?

[Jeeppler]

@msvechla

I used the Terraform AWS EKS module. The image is calculated by using the following snippet:

source: terraform-aws-eks/data.tf

data "aws_ami" "eks_worker" {
  filter {
    name   = "name"
    values = ["amazon-eks-node-*"]
  }

  most_recent = true
  owners      = ["602401143452"] # Amazon
}

[msvechla]

@Jeeppler perfect, we are using the same. Then re-running terraform and terminating the running instances should fix the issue. Thanks!

[Jeeppler]

@msvechla yes. I created a cluster from scratch. I did not try to change the instance type in a already existing cluster.