A method, a device, and a non-transitory storage medium are described in which a machine learning-based resource management service is provided. A network device obtains network and end device information, and uses machine learning to determine whether to adjust an auto-scaling rule pertaining to the provisioning of an application service. The network device generates a modified auto-scaling rule based on the analysis.