Online learning is one of the learning methods implemented by many educational institutions. One of the implementations is the online English Proficiency Test (EPT) with Moodle LMS as the examination platform. However, in practice, some cheating still occurs, and many examinees in the online EPT exam make it challenging to analyze the data and detect cheating. Knowing this problem, this paper proposes a log analyzer framework that could detect cheating in online EPT exams in real time using log data as the analysis data. While also implementing proactive forensic practices and digital forensics readiness in the system. The waterfall method was adopted to develop the log analyzer framework by developing the capability of data acquisition and replication, identifying features that can indicate cheating, implementing isolation forest machine learning models for cheating detection, and reporting the analysis results from data that can be used for further processing. The functional test of the system states that the log analyzer framework’s ability to acquire data, detect cheats, and create reports is functioning successfully. These functions demonstrate the log analyzer’s ability to strengthen the security of online EPT examinations.