Design And Implementation Of Cloud-Native Application Platform Based On Kubernetes

People's daily lives and software application services are increasingly inseparable.In order to respond to changing needs,it is becoming increasingly important for software development teams to quickly iterate.With the gradual development of microservices,DevOps,and container technologies,cloud native has emerged at the historic moment,providing solutions for rapid iteration of software application services.At present,Google has open sourced Kubernetes to help with container management,but mainly as a container orchestration tool,it cannot fully cover the entire process of cloud native applications from source code to release.This thesis implements a cloud-native application platform based on Kubernetes.Unlike Kubernetes,which focuses on container orchestration,the platform is application-centric,helping users to easily manage Kuberbetes-based applications.The platform provides an easy-to-use operation interface,while reducing the user's learning cost of using the cloud platform,it also provides support for the applications running on it from the four aspects of construction,deployment,monitoring,and performance.(1)Providing a build pipeline for application packaging,allowing applications to update continuously and quickly.(2)Providing blue-green deployment and gray-scale release deployment strategy support for application deployment,so that applications can be deployed and updated continuously.(3)Platform ingrates monitor system which allows users to intuitively view the resource consumption of the application.Meanwhile,it provides the alarm function for users,and informs users when abnormal conditions occur.(4)Providing a dynamic load balancing controller based on service quality for applications which adjusts the score of each node based on the response time and response error rate of each node to improve the overall service quality of the application cluster.(5)Providing auto scaling based on prediction and conservative constant for application clusters,allowing applications to scale horizontally and automatically based on load conditions,improving overall concurrent performance.Through testing,the platform meets expectations in terms of functionality,and the overall performance of applications running on the platform can be effectively improved through load balancer and auto scaling.