One of our Redshift clusters was getting stuck on certain operations. It caused a bottleneck on Provisioning API that delivers database credentials to transformations and other components.
Affected orchestration have been restarted. We're sorry for this inconvenience.