Riverside provides a single place to record, edit, and share conversations with the world. Creators and companies like Netflix, Marvel, and Shopify rely on Riverside to produce podcasts, video interviews, marketing videos, virtual conferences, internal communications, and more. With high-quality local recordings, AI transcriptions, and text-based editing, Riverside makes professional video creation achievable by anyone, anywhere.
As a high-load startup, the Riverside team leverages AWS EKS to provide secure, reliable, and scalable hosting of its workloads. The company engaged a team of DevOps experts from Automat-it and asked them to lead the modernization project, with the aim of replacing a third-party live call solution with Riverside’s applications.
The main considerations of the architecture were:
- Scalability to accommodate the increasing number of users talking in live voice chats.
- Cost efficiency to ensure optimal cloud spend.
- High performance for global users.
- High availability.
To accomplish the project, the joint Riverside/ Automat-it team built a multi-region infrastructure on AWS, with EKS clusters in six regions across the globe, using Karpenter as the node scaling mechanism and GitHub Actions/ArgoCD setup as the foundation of CI/CD workflow. The solution automatically purges unused resources, ensures optimal traffic routing with minimal latencies and jitter, and automatically routes traffic to a secondary region in case of failures, to minimize downtime for Riverside’s users.
Outcomes and benefits:
- The infrastructure uses Terraform as an IaC framework, and can be extended to additional regions when required.
- The CI/CD flow ensures smooth testing and deployment of the application changes across all EKS clusters.
- Using two AWS regions per geographical location ensures high availability, even in the worst outage cases.
