Refactor settings structure

[pehala]

• Rename kuadrant to service_protection
  • Move authorino and envoy under service_protection
• Rename mgc to control_plane
• Move kuadrant.enable to gateway_api
• Change how OpenShift are configured:
  • Introduce cluster configuration, which represent the cluster info where everything should run
    • By default is completely empty, which means that whatever server user is logged in will be used
    • This would be overridden if ran with MGC as the workload will be automatically moved to a spoke
  • All settings in service_protection section are only a projects on the cluster, no other configuration is allowed
  • control_plane.spokes and control_plane.hub (new settings) require a full cluster configuration, same as cluster
• Change openshift() and openshift2() fixtures to a module scope
  • When running future tests with MGC, the actual client is dependant on the hub gateway and spoke
  • This change means that backend is now also module scope, meaning slower tests

Current minimal configuration:

default:
  skip_cleanup: false
  dynaconf_merge: true
  gateway_api: true
  control_plane:
    hub:
      project: "multi-cluster-gateways"
    spokes:
      local-cluster: {}

[averevki]

I was thinking about control_plane.enabled: true flag instead of marking every mgc test (I'm open to creating PR after this is merged). What do you think?

[pehala]

I was thinking about control_plane.enabled: true flag instead of marking every mgc test (I'm open to creating PR after this is merged). What do you think?

I think we will use a different "smarter" way of marking them. I would like to keep the commandline switch