Run the staging AI-gateway in us-east4
In https://gitlab.com/gitlab-org/quality/quality-engineering/team-tasks/-/issues/2599 we've seen that the responses from Vertex are different when targeting us-east1
versus the responses in us-east4
that we're currently using in Production.
This would mean that for AI-gateway, we'd also be moving traffic from us-east1
to us-east4
in production when we want to roll out regional deployments.