Micro-service elastic scaling scheduling method and system oriented to time delay and resource utilization rate

The invention discloses a micro-service elastic scaling scheduling method and system oriented to time delay and resource utilization rate. The method comprises the following steps: carrying out real-time statistics and storing throughput and execution time of each API (Application Program Interface) corresponding to each micro-service of a call chain, the number of containers in a running state and a runnable state of each micro-service and a service dependency relationship in a database; giving a single-step prediction interval upper bound and a multi-step prediction interval upper bound of the throughput of each API facing the user micro-service; according to the single-step and multi-step prediction results and database information, calculating the number of container instances which need to be switched between a running state and a runnable state of each micro-service at the current moment and the number of container instances which need to be created or terminated; and selecting the weight of a Kubernete