Wait deployments Available instead of pods Ready

Change MetalLB controller-manager and webhook-server wait tasks to
check for Deployment Available condition instead of Pod Ready condition.
This ensures service endpoints are fully registered before applying
MetalLB CRs, preventing intermittent webhook errors:
"failed calling webhook...no endpoints available for service"

The Available condition on Deployments is more reliable than Pod Ready
for webhook availability, as it confirms the service is reachable.
This follows the same pattern used for NMstate webhook deployment.

Jira: OSPCIX-1095
Assisted-By: Claude Code/claude-4.5-sonnet
Signed-off-by: Harald Jensås <[email protected]>

Author: hjensas

roles/kustomize_deploy/tasks/install_operators.yml

@@ -230,39 +230,39 @@
         name: openshift_setup
         tasks_from: patch_dependent_operators_source.yml
 
-    - name: Wait for controller-manager pods
+    - name: Wait for controller-manager deployment
       kubernetes.core.k8s_info:
         kubeconfig: "{{ cifmw_openshift_kubeconfig }}"
-        kind: Pod
+        kind: Deployment
         namespace: metallb-system
         label_selectors:
           - control-plane = controller-manager
         wait: true
         wait_condition:
-          type: Ready
+          type: Available
           status: "True"
         wait_timeout: 300
       retries: 3
       delay: 60
-      register: _controller_manager_pods
-      until: _controller_manager_pods is success
+      register: _controller_manager_deployment
+      until: _controller_manager_deployment is success
 
-    - name: Wait for webhook-server pods
+    - name: Wait for webhook-server deployment
       kubernetes.core.k8s_info:
         kubeconfig: "{{ cifmw_openshift_kubeconfig }}"
-        kind: Pod
+        kind: Deployment
         namespace: metallb-system
         label_selectors:
           - component = webhook-server
         wait: true
         wait_condition:
-          type: Ready
+          type: Available
           status: "True"
         wait_timeout: 300
       retries: 3
       delay: 60
-      register: _webhook_server_pods
-      until: _webhook_server_pods is success
+      register: _webhook_server_deployment
+      until: _webhook_server_deployment is success
 
     - name: Wait until NMstate operator resources are deployed
       kubernetes.core.k8s_info: