Using multiple linear regression algorithm and design patterns to complete Kubernetes cluster configuration files

The problem of configuration of Kubernetes clusters is considered. Since cluster settings are made using YAML configuration files containing a large number of parameters, links to repositories (open and closed) and external data sources, it is quite easy to make a mistake that will incur significant costs in the future. When all the necessary data for the file is correctly prepared, it is necessary to arrange them correctly in accordance with the syntax of the YAML markup. The purpose of the work is to search for the optimal method of automatization of building a Kubernetes configuration file based on statistical data.The analysis of the Kubernetes configuration problem based on YAML files and the problem of interpreting statistical data into a structured YAML file is carried out and specific methods and approaches to solving these problems are proposed. The modified algorithm of multiple linear regression for working with the collected statistical data, the result of the output data of the algorithm and the flowchart of the pattern adapted for building YAML files are presented. The proposed approaches make it possible to use additional tools to work with Kubernetes test and production clusters, which reduces the complexity of developers' interaction with them and increases deployment speed and scalability. In addition, the described methods make it possible to simplify the administration of large networks and automate the process of creating configuration YAML files for popular software templates.