Kubernetes Ingress/Load-Balancing on AWS

7/3/2018

I'm used to using GKE for deploying kubernetes clusters, but now I've been tasked with deploying a load balanced cluster in AWS using HAproxy(or something similar) so I'm a little lost..

Currently I have 1 master node, 3 worker nodes, and a MySQL database node in a private subnet, and one node with an external ip address in a public subnet that I plan on using as an external load balancer. I'm also currently using Weave Cloud deployed via helm.

My question is:

Should I install HA proxy first, then deploy my applications onto the cluster, then figure out all of the routing? or vice versa? and I know that there are load balancers that can be deployed via helm, but wouldn't my cluster have to be in a public subnet for that to work? I don't want to expose the entire cluster to the internet so hopefully someone has some insight.

Thanks!

-- Coleman Word
amazon-web-services
kubernetes

0 Answers