Computer resource scheduling system based on reinforcement learning under micro-service architecture

The invention relates to the technical field of cloud service instance elastic scaling, in particular to a computer resource scheduling system based on reinforcement learning under a micro-service architecture, and each instance monitored by the computer resource scheduling system is provided with a resource capacity vertical scaling module; the computer resource scheduling system comprises a grouping device, a monitor, a decision maker, an actuator and a parameter adjuster. The monitor is used for monitoring the latest state information of each instance; the state information comprises the CPU resource utilization rate and the average response time of the request; the grouping device is used for logically grouping each instance according to a preset grouping requirement before each time of stretching; the grouping requirement is that the instances are divided into a plurality of Agent groups, a plurality of tentative groups and a threshold group; wherein the number of the instances in each Agent group is a p