Achieving Continuous Delivery of Immutable Containerized Microservices with Mesos/Marathon

Abstract

In the recent years, DevOps methodologies have been introduced to extend the traditional agile principles which have brought up on us a paradigm shift in migrating applications towards a cloud-native architecture. Today, microservices, containers, and Continuous Integration/Continuous Delivery have become critical to any organization’s transformation journey towards developing lean artifacts and dealing with the growing demand of pushing new features, iterating rapidly to keep the customers happy. Traditionally, applications have been packaged and delivered in virtual machines. But, with the adoption of microservices architectures, containerized applications are becoming the standard way to deploy services to production. Thanks to container orchestration tools like Marathon, containers can now be deployed and monitored at scale with ease. Microservices and Containers along with Container Orchestration tools disrupt and redefine DevOps, especially the delivery pipeline.

This Master’s thesis project focuses on deploying highly scalable microservices packed as immutable containers onto a Mesos cluster using a container orchestrating framework called Marathon. This is achieved by implementing a CI/CD pipeline and bringing in to play some of the greatest and latest practices and tools like Docker, Terraform, Jenkins, Consul, Vault, Prometheus, etc. The thesis is aimed to showcase why we need to design systems around microservices architecture, packaging cloud-native applications into containers, service discovery and many other latest trends within the DevOps realm that contribute to the continuous delivery pipeline. At BetterDoctor Inc., it is observed that this project improved the avg. release cycle, increased team members’ productivity and collaboration, reduced infrastructure costs and deployment failure rates. With the CD pipeline in place along with container orchestration tools it has been observed that the organisation could achieve Hyperscale computing as and when business demands.