When cloud costs grow faster than expected: bringing spend under control
e-Spirit AG is a German software company with a leading enterprise CMS platform (FirstSpirit), acquired by Crownpeak in 2022. The company operates infrastructure across 4 international locations. Cloud costs on AWS had grown over the years, uncoordinated, with historically accumulated structures at each location. A unified, cost-efficient infrastructure had become a strategic necessity.
CCsolutions took on the analysis, redesign, and migration of the entire AWS infrastructure to Kubernetes. The goal: lower costs, more visibility, faster delivery.
Challenge: heterogeneous infrastructure across 4 locations
The problem was not a single inefficient resource. It was the sum of all inefficiencies:
- Different infrastructure setups at each location, with no normalization
- Manually managed resources without a unified tooling strategy
- Cloud costs without transparency: unclear which workloads cost how much
- Delivery pipelines varied by team and location
- No unified security and compliance baseline
The objective: unified infrastructure. One standard for all locations, automated, auditable, and cost-optimized.
Solution: Kubernetes on AWS with Terraform, Helm, and GitLab CI/CD
CCsolutions developed a unified infrastructure architecture and rolled it out across all 4 locations.
Kubernetes as the common platform: All workloads were migrated to Kubernetes on AWS. Container-based deployments replaced varied VM setups. Resources can be scaled precisely, and idle capacity is reduced.
Terraform for all locations: The entire AWS infrastructure, including VPCs, subnets, EKS clusters, and IAM policies, is managed through Terraform modules. A shared module repository adapted for each location. Infrastructure drift becomes visible and correctable.
Helm Charts for standardized deployments: Application deployments via Helm Charts, defined once and rolled out consistently across all environments. No more divergence between locations.
GitLab CI/CD: Full deployment automation from feature branch to production, with automated tests, security scans, and controlled rollouts.
FinOps visibility: A clear tagging strategy for all AWS resources. Costs are visible per workload, per team, and per location. Idle resources are automatically identified and eliminated.
Results: 45% cost reduction, faster delivery
The results were measurable:
- 45% reduction in AWS cloud costs: through right-sizing, Spot Instances, and elimination of unused resources
- Normalized infrastructure: All 4 locations run on the same Kubernetes stack
- Faster delivery: CI/CD pipelines reduce time-to-deploy from hours to minutes
- Improved security: Consistent IAM policies and network policies across all clusters
- Compliance: Full auditability of all infrastructure changes via Terraform
Technology stack
| Area | Technology | |---|---| | Cloud | Amazon Web Services (AWS / EKS) | | Infrastructure as Code | Terraform | | Package Management | Helm Charts | | CI/CD | GitLab CI/CD | | Container Runtime | Kubernetes | | Observability | CloudWatch + Grafana |
Summary
The 45% cost reduction at e-Spirit is the direct result of a consistent strategy: normalize infrastructure, make costs visible, and systematically eliminate idle capacity. Kubernetes on AWS is the enabler here, not the goal in itself.
If your cloud costs have grown and you want visibility into what is costing how much, talk to us.