chore: check if AIGateway CRD exists in the cluster before creating a count provider

[pmalek]

What this PR does / why we need it:

• Prevent users being confused by the error message being logged when AIGateway count provider fails to get created

  Related issue: Controller wants to watch incorrect resource kind "aigatewaies" #297

• Additionally this creates a CRD guard so that when user requests AIGateway controller but the CRD is not in place then controller exits with an error.

• This also fixes the aigateway reports in telemetry. For some reason (probably due to the incorrect pluralized version aigatewies used in AIGateway GVR).

  The reports now include the AIGateway entry:

  report: <14>signal=gateway-operator-start;hn=P-Maek-MBP;uptime=0;v=NOT_SET;k8s_aigateways_count=1;k8s_controlplanes_count=0;k8s_controlplanes_requested_replicas_count=0;k8s_dataplanes_count=0;k8s_dataplanes_requested_replicas_count=0;k8s_gatewayclasses_count=1;k8s_gateways_count=0;k8s_httproutes_count=0;k8s_nodes_count=1;k8s_pods_count=14;k8s_referencegrants_count=0;k8s_services_count=4;k8s_standalone_controlplanes_count=0;k8s_standalone_dataplanes_count=0;k8s_arch=linux/arm64;k8s_provider=kind;k8sv=v1.30.0;k8sv_semver=v1.30.0;
  

Special notes for your reviewer:

For whatever reason the tests do not allow the full suite to pass as it used to. It's most likely relate to the pluralising issue in k8s: kubernetes/client-go#1082

[randmonkey]

Does this PR need a changelog entry?

[pmalek]

Does this PR need a changelog entry?

@randmonkey No need for changelog, this is not user facing.