Lessons Learned Migrating an On-Prem App to AWS EKS w/ Fargate
summary
Migrated a small Java web app to AWS EKS with Fargate and documents what actually changed and why it worked and where it struggled. Uses eksctl and scripted installs to reproduce clusters, learns that profiles and namespaces are required for Fargate, and that ALB role and cluster/VPC mismatches cause long debugging. Finds workload sizing matters (memory misestimates stop pods), horizontal scaling on Fargate is slower (Fargate scaling ~minutes), and Spot or Lambda can cut long‑term costs.
AI-written summary · about
chapters (14)
- 0:00Migration journey intro
- 1:12Speaker context
- 2:16EKS and Fargate
- 5:06App overview and goals
- 6:09Lessons learned
- 6:30Config automation
- 11:30Profiles troubleshooting
- 13:15Application Load Balancer configuration pitfalls
- 15:01Day two Fargate ops
- 15:43Workload sizing
- 17:36Horizontal scaling speed
- 19:43Cost control
- 21:51Cost savings alternatives
- 22:55Final recommendations
related talks
27mLeveraging AWS Fargate for running cost-friendly containerized application
49mHardware to AWS, EC2 to EKS: Two Years, Two Epic Migrations
21mDeploying Kubernetes at scale
46mArchitecting for Success and Runway
29mMigrating to AWS - Lessons Learned
38mThe Reality of Porting Kubernetes Apps Across Cloud Providers
40mServerless Compute at the Heart of Your EDA
39mLessons Learned From Maintaining Continuous Delivery While Migrating From a Static Infrastructure to Kubernetes [B]