Abstract: Jobs are scheduled per user across tenants in a multi-tenant environment. When a tenant is serviced, scheduling information is stored that indicates a servicing time of a user that remains unserviced after servicing the tenant. Each time a scheduler begins to schedule the servicing of tenants it obtains the list of the tenants. The scheduler sorts the tenants using the servicing information obtained from each of the tenants. The tenant that has the oldest unserviced user is serviced first. The scheduler starts servicing the first tenant in the sorted list and services as many jobs for as many users for that tenant based on the available processing resources. When the limit of servicing users for the tenant is reached and one or more users remain unserviced, the time for an unserviced user is stored in the servicing information.