fix: SSH provisioning for existing node pools when AppendRKEProvisionUserData is false

[LiviusP]

Description

Fixes SSH provisioning issue for existing node pools when AppendRKEProvisionUserData is false.

Problem

• Existing node pools failed to deploy properly because SSH provisioning via Rancher wasn't working
• When AppendRKEProvisionUserData was false (default), SSH user configuration wasn't added to cloud-init
• This caused the deployment to end with "Custom install script was sent via userdata, provisioning complete..." but SSH access failed

Solution

• Always add SSH user configuration to cloud-init when RKE provisioning is involved
• Ensures proper SSH access for both new and existing node pools
• Maintains backward compatibility

Changes

• Modified GetFinalUserData() to check for RKEProvisionUserData presence
• Updated tests to reflect new expected behavior
• Updated changelog for version 7.0.9

Testing

• ✅ All existing tests pass
• ✅ New behavior tested with TestGetFinalUserData* tests
• ✅ Backward compatibility maintained

Related Issue

Closes #128

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• Tests updated
• [] Documentation updated

Checklist

• PR name added as appropriate (e.g. feat:/fix:/doc:/test:/refactor:)
• Tests added or updated
• Documentation updated
• Sonar Cloud Scan
• Changelog updated and version incremented (label: upcoming release)
• Github Issue linked if any
• Jira task updated