Remove the tagging functionality of the VSphereFailureDomain API

[srm09]

/kind feature

Describe the solution you'd like Traditionally VI admins have been responsible for tagging and managing the vCenter resources. Having CAPV tag resources out of band is a departure from this behavior. From the feedback we received from our users, this is a functionality that they do not want to expose to CAPV, but have the VI admins perform resource tagging as a pre-requisite.

Anything else you would like to add: Deprecate the functionality via the following steps:

• Introduce the deprecation notice on the API type
• Remove the controller logic for resource tagging

Environment:

• Cluster-api-provider-vsphere version:
• Kubernetes version: (use kubectl version):
• OS (e.g. from /etc/os-release):

[sbueringer]

Discussed with Sagar, some additional information:

• autoConfigure takes care of creating the tags/categories, but does not delete them on cleanup
• once the tag/category is created, it can be used by other resources in VC, so deleting it might not lead to the best outcome

[sbueringer]

/retitle Remove the tagging functionality of the VSphereFailureDomain

We'll use this issue to track the deprecation and removal. PR for deprecation is open: https://github.com/kubernetes-sigs/cluster-api-provider-vsphere/pull/2103

/kind release-blocking (for the deprecation part)

[sbueringer]

/retitle Remove the tagging functionality of the VSphereFailureDomain API

[randomvariable]

/lifecycle frozen

[sbueringer]

/kind api-change

[sbueringer]

/remove-kind release-blocking

[sbueringer]

Note: We should also update the failure domain proposal to simplify it once the auto configure field was removed

[sbueringer]

Field is now deprecated, will be removed with v1beta2