Cloud computing has become a cornerstone of modern information systems due to its scalability and flexibility. The ability to allocate resources on demand enables quicker responses in managing processing power, data storage, and network infrastructure. Many organizations still rely on manual oversight for system maintenance and management. As a result, when errors occur, human intervention is often required in real time leading to increased costs and delays in problem resolution. With the sheer volume of system logs, which can reach millions per second, manual monitoring becomes impractical. A promising solution to this challenge is proactive anomaly detection, trained on log data, leveraging advanced techniques such as machine learning (ML) methods and natural language processing (NLP). However, applying artificial intelligence (AI) in this context presents its own challenges, particularly in hyperparameter tunings. This study investigates the use of metaheuristic optimization for hyperparameter tuning in a classification model designed for error detection in cloud computing logs. A modified metaheuristic has also been introduced specifically to meet the needs of this researcher based on the salp swarm algorithm (SSA). Evaluations on real world datasets have shown promising results, with an accuracy as high as 0.985199 in the best case scenario.