Migrating from kOps to EKS with Minimal Developer Involvement
summary
A platform team migrates eight kOps clusters to Amazon EKS by copying deployments and routing traffic through a custom EKS migration proxy built on Skipper so services switch clusters without downtime and roll back instantly if needed. An in‑house deployer Smutje decides per-service targets, enabling migration of about 266 microservices across environments in rounds that average three months each. Node provisioning moves to Carpenter, runs ~90% spot instances, and yields roughly 25% cost savings while removing control‑plane maintenance.
AI-written summary · about
chapters (8)
- 0:00Introduction and team overview
- 1:59State before migration
- 12:24Network architecture and migration proxy
- 14:51Migration process and tooling
- 18:00Post-migration cluster overview
- 24:46Challenges and learnings
- 29:44Key takeaways and future migrations
- 32:14Q&A session
related talks
49mHardware to AWS, EC2 to EKS: Two Years, Two Epic Migrations
46mMigrating from VMs to K8s
23mParadox's EKS Migration
43mReplatform to Amazon EKS
09mHow migrating from ECS to EKS supercharged our infrastructure
29mThe Frugal k8s Engineer: Strategies for Amazon EKS
29mMigrating 20K nodes and scaling to production on EKS & MSK in 8 months
01h 02mLearnings from the Enterprise-level Kubernetes transformation at Elkjøp Nordic