Fix flaky upstream-authority-vault and upstream-authority-ejbca integration tests

[amartinezfayo]

The upstream-authority-vault and upstream-authority-ejbca integration test suites were intermittently failing with error: no matching resources found when running kubectl wait pods -l app=spire-server --for=condition=Ready.

After kubectl apply, Kubernetes creates the Deployment object first, then the Deployment controller creates the ReplicaSet, which then creates the Pod. If kubectl wait runs before any pods with the label selector exist, it exits with an error immediately. With set -e active in the test scripts, this causes the step to fail.

This PR replaces the four kubectl wait pods ... --for=condition=Ready calls in upstream-authority-vault/02-deploy-spire-and-verify-auth and the one in upstream-authority-ejbca/02-deploy-spire with kubectl rollout status deployment/spire-server, which operates on the Deployment resource directly and correctly handles the window between kubectl apply and pod creation.

Example of failure: https://github.com/spiffe/spire/actions/runs/24039273578/job/70107728628?pr=6833