feat: upgrade to the latest version of mcr

[OlegErshov]

On-behalf-of: SAP [email protected]
This pr is related to #225
Changes:

1. Upgraded to the latest the mcr version which has a breaking change in the providers setup. Now an apiexport provider requires:

• kubeconfig which points to the workspace where the apiExportEndpointSlice is located
• name of the apiExportEndpointSlice to discover it dynamically
  Also, now all mcr providers implement Runnable interface and therefore they will run with a manager and there's no need to run them separately

2. Refactored kcp api imports to follow the same naming pattern

[aaronschweig]

Looks good to me - can you maybe link the PR that changes the kubeconfig generation to point to the workspace and the helm chart changes before merging?

[OlegErshov]

I've implemented this solution without changing the operator's kubeconfig, it still points to the virtual workspace host because we need this url for allClient creation and operations across all the involved workspaces. Instead of changing the kubeconfig I've provided a function which creates a config referencing the platform-mesh-system workspace to initialize the apiexport provider. Also after the provider has discovered the apiExportEndpointSlice resource, the kubeconfig becomes useless for the provider and I thought that having virtual workspace url in the kubeconfig would be more convenient for the development because virtual url is very useful

If we want to have the original kubeconfig referencing the platform-mesh-system workspace, I could adjust the approach and fetch the apiExportEndpointSlice resource where it's needed