Zonga, a local streaming service that would provide music lovers unlimited access to over 36 million songs in high quality was in full growth mode back in 2018. Soon enough, the development team had to improve streaming and conversion speed while ensuring a scalable infrastructure that would accommodate more users and more music libraries. What they didn’t know is that they could also reduce costs by being among the early adopters of Kubernetes in their industry.
Following-up on our audit, it was clear that migrating to a public cloud provider and Kubernetes was the best response to Zonga’s scalability needs. We started by migrating the entire infrastructure to Amazon Web Services and deployed Kubernetes to ensure scalability and a flexible environment. In order to reduce latency and increase customer satisfaction, as the closest AWS Region was Frankfurt, Germany, we decided to build a custom CDN network.
Here are the key steps in our migration process:
- Migration documentation
- Cloud infrastructure setup using Terraform
- Kubernetes Deployment and configuration
- Database Migration on AWS
- Kubernetes Deployment
- Application containerisation
- Kubernetes configuration
- Application migration
- Move logging to the ELK stack
- Move CI/CD pipeline to Jenkins
- Monitoring using Prometheus/Grafana.
The team behind the project has been essential to the success of the new platform architecture. Our colleagues Andrei Petrescu and Andrei Manea have joined forces with Zsolt Tasnadi, Senior Technical Consultant at Zonga to prepare and perform the migration process in the best possible way—fast, with zero downtime and without interfering with the code.
“Working with CloudHero was like expanding our team of experts in a matter of days. Their expertise and passion for Kubernetes consolidated the confidence that we teamed up with the right partners to support us in reaching our goal—fast, reliable, and professional, just the way it should be.”, said Zonga Senior Technical Consultant, Zsolt Tasnadi.
Migrating to Amazon Web Services Cloud and Kubernetes combo has reduced hosting costs by 75%. More specifically, transcoding 100k mp4 files into 4 different bitrates for under 100$ using AWS spot instances was now possible in real-time.
The migration process could not have been performed seamlessly if we hadn’t paired with a team of highly skilled developers with a keen interest in learning and understanding Kubernetes. The communication and trust level enabled us to perform the migration process just two months, with zero downtime. The AWS and Kubernetes proved to be the right solution for Zonga to achieve infrastructure scalability, performance, and significant cost reduction.