This article examines the problems of distributed algorithms and proposes their solution using temporal analysis. There are many things that can go wrong in distributed systems that can cause the system to crash. The solution to this is to build a system that can withstand the problems that arise during its operation. It turns out that having an algorithm capable of reaching consensus is extremely important for systems that want to function properly despite network failures. Although consensus is omitted in performance-oriented systems, they still rely heavily on systems that implement consensus algorithms for them (such as Zookeeper, etc.) to handle the consensus-reduced task, while at the same time having some weaker consistency model. In turn, the algorithms available today have several problems, the solution of which will significantly increase the performance of the algorithms and, as a result, the systems that use them. This article discusses the problems that arise in existing implementations and presents a data analysis technique and model for solving one of the algorithm problem.