The increasing integration of artificial intelligence (AI) into everyday tasks has driven demand for high-performance software to an unprecedented level. This has resulted in an exponential growth in energy consumption by software systems. While extensive research has been done focusing on measuring and evaluating the energy consumption of machine learning and AI systems, particularly in Python, a significant gap remains in tools and methodologies for assessing energy usage in legacy systems. Legacy software systems continue to underpin critical operations across industries such as finance, government, and healthcare, yet their energy consumption remains largely unmeasured and unoptimized. As organisations face mounting pressure to improve sustainability and reduce operational costs, the lack of dedicated tools for quantifying the energy usage of legacy codebases, particularly those written in languages like COBOL, has become a significant barrier. In this paper, we discuss this critical challenge and specifically focus on the legacy language of COBOL. In particular, we present \textit{COBJoules} as one of the first tools designed to measure and analyse energy consumption in COBOL environments. The tool was evaluated using benchmark tasks from Rosetta Code, demonstrating its effectiveness in quantifying energy consumption patterns for COBOL. The tool and all related artifacts can be found at https://github.com/ethicode4-ops/Cobjoules.