During the Russia 2018 World Cup, the company faced extremely high expenses in infrastructure and serious availability issues that severely affected the service provided to its customers.
We designed and implemented a scalable and highly available infrastructure in AWS, which guaranteed the service availability, while achieving a cost reduction of 72%.
As we had a short time to implement the solution, we assembled a five people team to handle the different issues in parallel. The platform was developed in Ruby on Rails and, initially, it was manually deployed and managed inside a VPS in Linode.
The first thing we did was to design the AWS infrastructure that would meet the customer’s costs and availability expectations. After that, we containerized the application with Docker and deployed it in a Kubernetes cluster implemented with EKS.
To simplify the application management in production, we also implemented continuous deployment with rolling-updates so we could deploy each new version of the application in production quickly and without downtime. At the same time, we configured the infrastructure in a way that, if the demand increased new servers were created automatically to distribute the load, without any impact on the application’s performance. In much the same way, when the user demand decreased, the new servers were removed automatically to reduce costs.
The infrastructure included several dashboards with information regarding events and resources consumption in real time.