cortex
cortex copied to clipboard
Extend the IP address pool for the Cortex cluster
Description
At bigger scales, the cluster runs out of IPs for its nodes/pods.
Solutions
-
Use "Custom CNI networking", which basically adds another CIDR block to the VPC and thus allows a greater number of pods and greater number of nodes (as the pods will no longer get IPs from the default block). More on that here: https://docs.aws.amazon.com/eks/latest/userguide/cni-custom-network.html. The issue when trying to follow the instructions in this tutorial can be found here: https://weave-community.slack.com/archives/CAYBZBWGL/p1626554301154100.
-
Find out a way to restructure the subnet mask sizes so that the nodes/pods will get more private IPs. Maybe we don't need to have that many public IPs. More on that here: https://eksctl.io/usage/vpc-networking/