This almost works for getting things going, first installation via kots
almost works, but ws-daemon's complain about the k8s mapping / containerd
due to changes in the amazonlinux AMI vs GitPod's.

End of the readme includes the yaml snippet to run to redeploy the stack
with the correct containerdRuntimeDir

Also the AWS ALB can't seem to find the proxy node to reliably deploy
the complete end to end session from the public IP to the internal service

Next step to investigate would be something like:
https://kubernetes.io/docs/reference/access-authn-authz/authorization/#checking-api-access

Also to cleanup the eks-cluster.yaml a little more universal - eksctl
suports a `prebootstrap command`:
https://eksctl.io/usage/schema/#managedNodeGroups-preBootstrapCommands

Having the prebootstrap command provide the following execution might
get a decent work around to keep EKSCTL's launch template that auto detects
AWS AMI by region/resources, which we lose when override the BootstrapCommands
in order to enable containerd:

```
cat <<-EOF > /etc/profile.d/bootstrap.sh
export CONTAINER_RUNTIME="containerd"
export KUBELET_EXTRA_ARGS="--runtime-request-timeout 600s"
EOF

sed -i '/^set -o errexit/a\\nsource /etc/profile.d/bootstrap.sh' /etc/eks/bootstrap.sh
sed -i 's/KUBELET_EXTRA_ARGS=$2/KUBELET_EXTRA_ARGS="$2 $KUBELET_EXTRA_ARGS"/' /etc/eks/bootstrap.sh
```

This branch also shows how to append node labels via the template instead
of the userdata script itself, which will be much more robust going forward
and is one less reason to use the overridebootstrap option.