Feedback type automatic capacity expansion and contraction method and system based on cluster performance

Abstract

The present invention provides a feedback type automatic capacity expansion and contraction method and system based on cluster performance, and belongs to the technical field of communication. The processing method comprises: obtaining parameters of each node and container, collating and obtaining performance indicators of the entire cluster, and storing the performance indicators, wherein the storage format is a timestamp+json format; scoring the collected performance indicators by using a determination model, finally determining whether to perform capacity expansion and contraction, and giving a capacity expansion and contraction scheme; generating a capacity expansion and contraction configuration template according to the capacity expansion and contraction scheme, and issuing a capacity expansion and contraction command; and obtaining a capacity expansion and contraction execution result returned by the cluster. The processing system comprises: a performance collection subsystem and a capacity expansion and contraction determination and execution subsystem. According to the method and system provided by the present invention, the performance of a specific container can be perceived, and the overall performance of the host and the cluster of the container can be perceived to automatically complete capacity expansion and contraction of the container; and the problems that theexisting capacity expansion and contraction method has poor pertinence and a slow response speed are solved.

Description

technical field [0001] The invention belongs to the technical field of communication, and in particular relates to a feedback type automatic expansion and contraction method and system based on cluster performance. Background technique [0002] In actual production systems, we often encounter scenarios where a service needs to be expanded, or scenarios where the number of service instances needs to be reduced due to resource constraints or workload reduction. This situation is especially prominent in microservice scenarios based on container technology. The reason is that microservices divide a granular application system into different service modules, for example: divide a web application into five major service modules (1, load balancing module nginx 2, web server module 3, background database module 4, Cache module 5, session persistence module), each module consists of several containers with the same function (corresponding to pods of kubernetes) to form a load balanc...

AI Technical Summary

Problems solved by technology

[0005] For method 1, using external test tools to measure the response speed of the system and manually expanding and shrinking capacity based on this has the following disadvantages: First, it can only roughly grasp the overall response speed of the system, and cannot perceive the pressure and utilization of each service module. Carry out targeted expansion processing, and the second is that the manual expansion method cannot quickly respond to the expansion needs of the business

[0006] For method 2, use the HPA (Horizontal Pod Autoscaler) function of kube

Images