As of November 2019, we switched over to auto-scalable Kubernetes and tagged along the underlying infrastructure to GCP. Ramping more workers of an existing scriptworker simply means scaling some configurations. You can find out more information on this here