From Atlas to EKS: How to Stay Sane — and Save 45% — During a Large‑Scale MongoDB Migration

Cloud repatriation is on the rise — but migrating stateful, mission‑critical systems like MongoDB is far from simple. In this talk, I present a real‑world case study of moving seven production MongoDB clusters from MongoDB Atlas to AWS EKS, powered entirely by Percona’s open‑source ecosystem.

We’ll explore the challenges of handling heterogeneous workloads (including multi‑terabyte sharded clusters), implementing online migrations with near‑zero downtime, and building safe roll‑out/roll‑back procedures. I’ll also cover performance tuning on Kubernetes and how we used GitOps principles with ArgoCD to keep deployments predictable and repeatable.

Beyond the technical journey, I’ll share the business impact of this migration: by moving to a self‑managed Percona‑based platform on EKS, we reduced operational costs by approximately 45%, while improving transparency, flexibility, and long‑term control.

The session dives into the practical use of Percona Operator for MongoDB, Percona Server for MongoDB, PMM, MongoSync, and Kafka — highlighting what worked, what didn’t, the hidden “gotchas,” and the lessons learned that can help you navigate your own cloud‑exit or modernization journey (with your sanity intact).