Add integration coverage for Route53 record retry-cap requeue

Summary

Add envtest or integration-level coverage for the retry-cap path introduced in PR #460. When VPC endpoint validation fails after the VPC endpoint is ready but the Route 53 record is not ready, reconciliation should suppress that validation error and return a fixed one-minute requeue delay.

Rationale

The existing unit tests cover the Route 53 tag-condition cache gate and condition persistence. The retry-cap branch is in VpcEndpointReconciler.Reconcile, and reliably asserting its returned ctrl.Result requires a reconciler-level environment.

Affected area

• controllers/vpcendpoint/vpcendpoint_controller.go
• New or existing envtest/integration test infrastructure for the VpcEndpoint reconciler

Acceptance criteria

• Set up a VpcEndpoint state where AWSVpcEndpointReady=True, AWSRoute53RecordReady is not True, and resource validation returns an error.
• Invoke reconciliation through an envtest or integration-level test.
• Assert reconciliation returns no error.
• Assert ctrl.Result.RequeueAfter == time.Minute.
• Assert the behavior remains scoped to this condition combination; unrelated validation errors must continue to be returned normally.

Backlinks

• PR #460: https://github.com/openshift/aws-vpce-operator/pull/460
• Review discussion: https://github.com/openshift/aws-vpce-operator/pull/460#discussion_r3679152330

Requested by @dustman9000.