The cloud migration met expectations in agility. But the monthly AWS invoice was growing faster than the business. FinOps helped us reduce costs by 40%.
Where the Money Was Leaking¶
- Dev environments running 24/7 (utilized only 8h/day)
- Oversized instances — t2.xlarge where t2.medium was sufficient
- Orphaned EBS volumes and snapshots
- Unoptimized S3 lifecycle policies
- NAT Gateway data transfer costs
Quick Wins¶
Scheduling: We automatically shut down dev and staging environments at 7 PM and start them at 7 AM. Weekends off. Savings: 65% on non-prod compute.
Right-sizing: AWS Cost Explorer and CloudWatch metrics revealed that 70% of instances were oversized. Downsizing with no impact on performance.
Reserved Instances: For stable workloads, we purchased 1-year RIs. 40% discount compared to on-demand.
FinOps Culture¶
The most important change: every team sees their costs. Tagging strategy — every resource has tags for team, project, environment. Monthly cost reviews with team leads. Costs are part of architectural decisions.
Tools¶
AWS Cost Explorer, AWS Budgets with alerts, Spot.io for spot instance management. Considering Kubecost for Kubernetes cost allocation per namespace/service.
Cloud Isn’t Cheap — Without FinOps¶
Cloud is cheap if you manage it correctly. FinOps practices saved us thousands of euros per month with no impact on performance or availability.
Need help with implementation?
Our experts can help with design, implementation, and operations. From architecture to production.
Contact us