only managed databases (RDS) • no named servers / ssh access anymore • all dockerized • use CodeDeploy / CodePipeline Backend Plattform Finance Tools End-user Website Refactor setup for AWS best practices Our journey: Strategy
replication 2. Bring up duplicated instances at AWS 3. Put 5% traffic to AWS to iron out bugs 4. Put 50% traffic to AWS to A/B test 5. Put 100% traffic to AWS 6. Setup 2nd region in AWS for disaster backup and re-establish master-master replication 7. Return physical server to hoster
Consultant / AWS Solution Architect • Sandbox with full admin to all developers • Get buy-in from colleagues • Collect their concerns • Hands-on Hackathon, plattform engineers help them out • Calculate costs together • A/B testing datacenters • Route53 round robin and sticky sessions don’t play well • Branch of at loadbalancer of physical hoster