Novel tools and techniques for controlling workloads in a multi-tenant environment. Some such tools provide a queue controller that can control workflow processing across systems, work (provisioning engines, computing clusters, and/or physical data centers. In an aspect, a queue controller can determine the status of each work request based on one or more attributes, such as the workflow type, the systems affected by (and/or involved with) the workflow, information about the tenant requesting the workflow, the job type, and/or the like. In another aspect, a queue controller can be policy-based, such that policies can be configured for one or more of these attributes, and the attribute(s) of an individual request can be analyzed against one or more applicable policies to determine the status of the request. Based on this status, the requested work can be scheduled.