Migrating Massive Aurora and MySQL Databases to Vitess Kubernetes Clusters with Near-Zero Downtime

Day 1 | 16:35 | 00:30 | UA2.114 (Baudoux) | Matthias Crauwels, Rohit Nayak


Note: I'm reworking this at the moment, some things won't work.

The stream isn't available yet! Check back at 16:35.

Vitess, a CNCF-graduated, open-source, MySQL-compatible database, is designed for massive scalability and cloud-native deployments. As one of the most robust solutions for managing large-scale database workloads, Vitess on Kubernetes powers some of the largest systems in the world.

Modern data teams face a growing challenge: scaling databases dynamically to handle surges, such as holiday sales or viral trends. For many, vertical scaling has reached its limits, leaving fundamental architectural changes as the only option. Vitess, with its built-in horizontal sharding capabilities, provides a seamless path for organizations to scale beyond the constraints of legacy MySQL or Aurora clusters. This talk explores how Vitess enables smooth migrations from Aurora and MySQL into Kubernetes clusters with near-zero downtime. We will begin with an overview of Vitess’ architecture, highlighting horizontal sharding and its advantages for scaling. Next, we will dive into the powerful workflows Vitess provides for data import and live production migrations, emphasizing their safety, efficiency, and reversibility—even for massive datasets.

To ground these concepts, we will share real-world examples of successful migrations from web-scale Aurora and legacy MySQL databases, including performance metrics, data sizes, and challenges encountered. Attendees will gain practical insights and a clear blueprint for transitioning large-scale database systems to Vitess, unlocking unprecedented scalability and operational agility.