CircleCI’s Technical Architecture Is Built for Scalability
Like a lot of startups, CircleCI started with a monolith, running on Amazon Web Services (AWS). But there was only so much that could be done to keep the monolith performing when by 2015 the company’s success required greater scalability. On this new episode of The New Stack Makers, TNS founder Alex Williams caught up with CircleCI’s Chief Technology Officer Rob Zuber in a live-streamed discussion to dig deeper into the architecture workings behind the organization.
CircleCI had built its own job scheduler in 2011 because there was nothing on the market that matched what they needed. By 2015, however, the company had made a conscious decision to use third-party tools instead of building its own, including the Cloud Native Computing Foundation’s Kubernetes container orchestrator, Docker containers, AWS, and HashiCorp’s Nomad scheduler. In managing Kuberenetes on AWS, Zuber noted that CircleCI has faced some issues as Kubernetes is not officially supported by AWS, but that the overall experience has been positive.
“In terms of managing our own infrastructure, what mattered to us was the ecosystem,” he said. “Kubernetes is a scheduler, but with it are all of these tools people have built to help us do our job. It was becoming pretty clear Kubernetes was a leader in terms of the amount of support it was getting.”
In This Edition:
1:40: Exploring the technical architecture and infrastructure behind CircleCI.
6:23: How CircleCI approached its transition from a monolithic infrastructure.
11:12: CircleCI’s use of Nomad and Kubernetes in its infrastructure.
18:07: How CircleCI manages its Kubernetes infrastructure on AWS.
21:35: The additional tools CircleCI is using to manage its infrastructure.
23:14: Exploring where the CircleCI 2.0 platform is heading.
The Cloud Native Computing Foundation is a sponsor of The New Stack.