[EKS] Support custom instance roles for node groups

[thedadams]

In order to support custom node instance roles for EKS clusters, a new dropdown should be added to the "Add nodegroup" section of the UI.

This dropdown should:

1. Get the available node instance roles from the AWS API to populate the dropdown.
2. Filter out the node instance roles that are prefixed with rancher-managed-
3. The field name should be nodeInstanceRole
4. The value passed to the Rancher API should be a string. If no role is selected for a nodegroup, then the empty string should be passed.

SURE-4536

[snasovich]

@nwmac @gaktive , is this something that can be implemented in 2.6.3? We're working on getting backend done so it's technically backend blocked, but hopefully some work can be started with the assumption this new field is added as per PR linked to backend issue.

[gaktive]

@snasovich do you have the backend ticket somewhere to refer to? Otherwise, what is the state of the backend if we can target this for 2.6.4?

[thedadams]

Backend issue: rancher/rancher#29071

[snasovich]

@gaktive , so far we're tentatively targeting this for 2.6.5. I've updated your milestone to match.

[gaktive]

@snasovich i see that the backend ticket no longer has a milestone now. If that's expected, I can place this ticket in 2.6.x.

[snasovich]

@gaktive @nwmac , we're working on getting backend changes for this done for 2.7.0 (not 100% sure we'll make it by deadline though). Is it possible to have UI changes ready for 2.7.0 release?
cc: @Sahota1225

[gaktive]

@snasovich is the ask to have UI ready before backend? I'll check if we can start work and then wait, though do let us know if any design changes are anticipated.

[snasovich]

@gaktive , the plan is to get backend changes merged by end of day 9/20/22. I'm not sure how this aligns with your deadlines, so if starting before backend is ready helps, that would be good.
@annablender , let us know if whatever is outlined in the description of this issue needs to be updated based on your work on backend.

[Josh-Diamond]

Looks like drone failed, not quite ready to-test

Restarting drone build now, moving to review until drone passes.

[Josh-Diamond]

Unable to restart drone - [perhaps lack of permissions?]

Requested assistance from UI team (internally)

[a-blender]

@Josh-Diamond Check if you've logged in to drone

[Josh-Diamond]

Ticket #4568 - Test Results - ✅

With Docker on a single-node instance:

Verified on rancher v2.7-0504ce63c1cad67b2c2be14a0325533b60902832-head:

1. Fresh install of rancher v2.7-head
2. Provision a downstream EKS cluster w/ custom instance role assigned for node group
3. Once active, navigate to Cluster Management > Edit Config > Node Instance Role
4. Verified - accurate custom role is seen and present; as expected

Screenshot: