In our vpc architecture, we have designed like one vpc with two availability zones, two public subnets , 2 private subnets , 2 private db subnets, internet gateway , route table association to pulic subnets, 2 nat gateway for 2 az's , route table association to private subnets, single eks cluster with single node group ( 2 nodes) on terraform. This is working fine. But the cost of NAT gateway is getting more. For the previous month we are getting bill for $700 on AWS for the whole EKS architecture we designed.
So we planned to create a two different nodegroups with single node where we use taints , tolerations & node selector for both dev & qa nodes.
'I want clarification regarding istio.'
How this istio can communicate with each node groups onto single EKS cluster?
Whether i need to run this istio on both node groups seperately or for single istio enabled for whole cluster is enough?
If i enable the launch template for node group , i m getting error like node group not attached to cluster. How to solve this error.
Is there any other way to reduce the billing on AWS for EKS?
we have enabled prometheus & grafana & loki for every pod running? if we are using different node group then we need to install on each node strictly or not.
