<![CDATA[E-learning is a new way to learning to use electronic media typically involving the internet. E-learning has several advantages where students can access subject matter at any time. One of the most popular e-learning applications is Moodle. In its implementation, Moodle can be done using moodlecloud, on-premise and self-hosted servers. The implementation of the service moodle e-learning in self-hosted or using cloud computing services would be ideal option. The advantages offered by cloud computing services are a strong consideration, where not required financing and do not need to spend in early financing server maintenance. Service selection and deployment model greatly affect the performance of the Moodle application and also the financing from resource used. Moodle requires a service with good performance to cope with the number of users who use it, this becomes a polemic for moodle e-learning service providers, which is how the ideal deployment model to implement moodle e-learning applications on cloud computing services with high levels of performance and scalability and also service financing used. The test is done by providing requests in the form of enrol activities, view courses and quiz on docker swarm clusters built on aws services ec2 t2.medium, t3.medium and t3a.medium services with user rates of 40, 80, 120. The test includes performance testing with test parameters, namely throughput, success ratio, latency, use of cpu and memory resources. Based on the results of testing and cluster analysis, ec2 t3.medium and t2.medium have good performance based on the throughput and latency test values for enrol activities, view courses and quizzes, based on the success ratio test scores, clusters on ec2 t2.medium have good performance with an error value of 0% on enrol activities, 0% on view course activities and 0.1% on quiz activities with 120 user requests.]]>
