Abstract: The present invention relates to a technical concept for updating software and performing load balancing for one virtual AI component (nginx) without service interruption based on a reduced system configuration. The non-disruptive software update system based on a container cluster according to an embodiment performs a software upgrade through a software patch of an AI component (nginx), but a software update processor for monitoring whether a service is stopped while performing the upgrade, Load balancing processing to distribute the load by replicating the application of the AI component (nginx) into a plurality of, and the number of components if the CPU usage observed in the replicated applications increases beyond the reference, the load balancing processing unit If the increase and the CPU usage is lower than the reference may include an auto scaling processor for reducing the number of the duplicated applications.