Abstract <p>This article presents a method for evaluating the correctness of distributed algorithms using chaos engineering methods to improve the efficiency of their testing. Modern research studies are analyzed in the field of testing distributed systems, such as federated learning systems for large language models (LLMs) and chaos engineering. The existing testing methods and tools based on chaos engineering are analyzed and their shortcomings are identified. As a result, a method for evaluating the correctness of distributed systems through testing using chaos engineering is developed, and vulnerability testing is conducted in open source projects, followed by a comparison with the existing methods. The obtained results confirm the effectiveness of the proposed method.</p>