add instance connection draining for NLBs

[heybronson]

This builds on #12902 to add graceful connection draining for associated targetGroups.

Today, one can define targetGroupArns in an instanceGroup spec to associate instances in a corresponding instanceGroup with. When instances in the instanceGroup are cycled via rolling-updates, those instances aren't first properly deregistered with the corresponding targetGroupArns, causing dropped connections.

This PR fixes this issue by ensuring that we're properly draining instances from their corresponding classic load balancers and NLB targetGroups before terminating the instance. This is a big improvement to the overall client experience when running rolling-updates on production clusters.

[k8s-ci-robot]

Hi @heybronson. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[heybronson]

cc @olemarkus @johngmyers

[johngmyers]

/ok-to-test

[rifelpet]

For reference, this deregistration occurs after the node has been drained of its non-daemonset pods but kube-proxy is still running, so this specifically helps services exposed via kube-proxy.

This doesn't help target groups targeting pod IPs (which can be gracefully deregistered via finalizers and the aws-load-balancer-controller during the drain process) nor does it help targeting pods listening on the hosts' nodePorts directly (since they will have been drained already and their connections disconnected)

[heybronson]

For reference, this deregistration occurs after the node has been drained of its non-daemonset pods

@rifelpet That's correct, this deregistration happens after node draining, so certain service configurations are already covered. However, this helps in scenarios where traffic is routed to daemonset pods running on a node, since those aren't drained. For example, Nginx-ingress in daemonset mode benefits from this change.

[heybronson]

/assign @johngmyers

[heybronson]

/assign @olemarkus

[heybronson]

friendly bump

[olemarkus]

/lgtm

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: hakman

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [hakman]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[heybronson]

@rifelpet can we potentially keep this in the 1.23 release?