Stabilize imported generic cluster e2e spec

Fix flakiness on imported cluster create against Prime Rancher and
wrap cluster manager tests with their Qase case IDs:
- accept Pending as a transient state alongside Provisioning/Waiting
- select provider column by .col-cluster-provider (formatter-derived)
- wait for cluster list visibility before masthead/list interactions
- extend cluster detail registration page wait to medium timeout
- disable Cypress retries on create to avoid stale clusters
- wrap eligible it() statements with qase() using the oldest matching
  Qase case ID per title

Author: izaac

cypress/e2e/po/lists/provisioning.cattle.io.cluster.po.ts

@@ -33,12 +33,13 @@ export default class ProvClusterListPo extends BaseResourceList {
   }
 
   provider(clusterName: string) {
-    return this.resourceTable().sortableTable().rowWithName(clusterName).column(3);
+    return this.resourceTable().sortableTable().rowWithName(clusterName)
+      .get('.col-cluster-provider');
   }
 
   providerSubType(clusterName: string) {
-    return this.resourceTable().sortableTable().rowWithName(clusterName).column(3)
-      .find(`.text-muted`);
+    return this.resourceTable().sortableTable().rowWithName(clusterName)
+      .get('.col-cluster-provider .text-muted');
   }
 
   machines(clusterName: string) {