Success Story
Simetrik migrates 58 applications from Amazon ECS to EKS in 3 months
At a glance
Simetrik is a Colombian fintech that automates payment data reconciliation for finance teams. It had dozens of services on Amazon ECS that it needed to move to Kubernetes to gain control and scalability. Craftech built a platform on Amazon EKS with three environments, GitOps with ArgoCD and full observability. In about 3 months it migrated 58 applications, 9 more than planned.
58
applications migrated to EKS
+9
more applications than planned
~3 months
to deliver the production infrastructure on EKS
Simetrik is a Colombian fintech with an AI-powered no-code platform that automates the reconciliation of data from multiple payment sources for finance, accounting and FP&A teams. It has around 300 employees and operates in Latin America.
Solution implemented
- 3 Amazon EKS clusters with Terraform and Terragrunt
- 58 services deployed in DEV and PRD, with Amazon ECR for images
- ArgoCD and Argo Diff Preview
- Harness pipelines integrated with EKS and ArgoCD
- Custom Helm chart with Doppler support
- Prometheus, Grafana Alloy, Loki and Grafana
- Atlantis to govern infrastructure changes
- Cost optimization with nightly shutdown, Spot instances and EKS Auto Mode
Business Challenge
Simetrik had dozens of services on Amazon ECS (49 at the start, later 58) that had to move to Kubernetes to gain control and scalability and to simplify operations. The project required three environments (DEV, CROSS and PRD), GitOps-based CI/CD, infrastructure as code and centralized observability, without disrupting operations and following the Well-Architected Framework.
Solution
Craftech built a new container platform on Amazon EKS for every environment: 3 clusters with Terraform and Terragrunt, 58 services deployed in DEV and PRD, and Amazon ECR for images. Continuous deployment runs on ArgoCD, with Argo Diff Preview, and the pipelines were modernized in Harness, integrated with EKS and ArgoCD. It developed a custom Helm chart with Doppler support, implemented full observability with Prometheus, Grafana Alloy, Loki and Grafana, and added Atlantis to govern infrastructure changes. Costs were optimized with nightly shutdown, Spot instances and EKS Auto Mode.
Outcomes
- 58 applications migrated to EKS, 9 more than planned
- Production infrastructure on EKS delivered in about 3 months
- Automated CI/CD with ArgoCD in DEV and PRD
- Production observability with Prometheus, Grafana Alloy, Loki and Grafana
- DEV, CROSS and PRD environments in separate accounts with Well-Architected practices
- Simetrik ready to execute the final traffic cutover to EKS on its own
Stack & technologies
