Cloud Migration — First Steps and Lessons from Our Mistakes

After years of on-premise operations, we decided to go cloud. Not because of hype, but out of necessity — clients want elasticity, multi-region availability, and faster provisioning. Here are our first experiences with AWS.

Why AWS¶

The largest ecosystem, the most services, the most documentation. Most of our clients are on AWS or planning a migration. And crucially: AWS has a region in Frankfurt — data residency for EU clients solved.

What We Underestimated¶

Networking. VPC design is critical and hard to change later. CIDR planning, subnet strategy, VPN/Direct Connect — this has to be right from the start.

IAM complexity. AWS IAM is powerful but incredibly complex. Policies, roles, instance profiles, cross-account access — the learning curve is steep.

Costs. Pay-as-you-go sounds great until the first invoice arrives. Data transfer, NAT Gateway, CloudWatch logs — hidden costs add up.

What Pleasantly Surprised Us¶

Managed services. RDS instead of self-managed PostgreSQL. ElastiCache instead of self-managed Redis. Operational burden dropped dramatically. Automatic backups, patching, multi-AZ — all without our effort.

Elasticity. Auto Scaling Groups for EC2, Aurora Serverless for databases. Scaling that simply didn’t exist on-premise.

Our Cloud Principles¶

• Infrastructure as Code from day one (Terraform)
• Encryption at rest and in transit — always
• Tagging strategy — who, why, how much does it cost
• Least privilege IAM — no admin credentials in code
• Multi-AZ for everything in production

Cloud Is a Journey, Not a Destination¶

First steps in the cloud are educational and sometimes painful. But the benefits — elasticity, managed services, global reach — are worth the investment.