[ AWS ]
Designing and operating your cloud-native platforms on AWS
EKS-First
Designing your cloud-native platforms on EKS
We build your EKS clusters to host e-commerce platforms, business applications and high-traffic back-office systems, leveraging AWS managed services without becoming locked in. The Kubernetes data plane is dealt with at the platform level, treated with the same rigour as an in-house product.
Move-to-Cloud
Modernising your legacy applications using AWS cloud-native architectures
Our CMS, Drupal and e-commerce experience enables us to migrate your existing platforms to AWS in stages, without any service disruption or the risks involved in big bang adoption. We don’t just move your problems elsewhere: we address them.
FinOps
Keep your AWS costs under control in the long term
Savings Plans, rightsizing and automated shutdowns in non-production environments: we make cost management part of an ongoing process of cost-benefit analysis. The best bill is one you don’t have to pay.
Our AWS expertise
- EKS and Kubernetes orchestration
We operate EKS using separate node pools (ingress, application, batch) to handle traffic spikes without oversizing the fixed pools. Cluster Autoscaler as standard; pod-based autoscaling depending on the customer’s requirements.
- Stateless refactoring and gradual migration
We manage cloud-to-cloud migrations and the modernisation of legacy platforms to AWS: audits, stateless refactoring, and gradual migration with no service disruption.
- Infrastructure as Code and reversibility
We standardise your AWS infrastructure using Terraform or OpenTofu, supplemented by AWS CDK for complex application logic. Our modules remain portable and compatible with other cloud platforms, driven by practicality rather than dogma.
- Stateless refactoring and gradual migration
We manage cloud-to-cloud migrations and the modernisation of legacy platforms to AWS: audits, stateless refactoring, and gradual migration with no service disruption.
- Observability and CloudWatch
We combine CloudWatch with the open-source ecosystem (Prometheus, Grafana, Loki, OpenTelemetry, Grafana Alloy) to provide 360° observability: logs, metrics and distributed tracing. Effective troubleshooting in production is achieved in advance, not in the heat of the moment.
- Advanced FinOps optimisation
In addition to Savings Plans and Database Savings Plans, we apply specific optimisation patterns: EventBridge automations to shut down EKS node groups outside of production hours, deletion of dormant environments, and forecasting of egress costs. FinOps is not an add-on, it is an everyday part of pipelines.
- Security and auditability
CloudTrail for auditing, least-privilege IAM, OIDC/Workload Identity to decouple static secrets, ECR with image scanning, and TFSec and Trivy integration in pipelines. AWS account security starts right from the pipeline.
The difference our AWS expertise makes
- AccorTMA and operation of the Group’s AWS platforms: continuous coverage, unified observability, AWS WAF, FinOps management of consolidated accounts.
- Yelloh! VillageAWS operation of the Strapi site factory deployed on Kubernetes, GitOps ArgoCD and end-to-end observability.
- ComexposiumMigration and optimisation of the events platform to EKS, with a Grafana/Loki/Thanos observability stack and a continuous FinOps programme.
- Forvis MazarsMulti-region managed services on AWS for an international audit and consulting firm: standardised operations, unified observability and coordinated version upgrades.
- French Red CrossSecure AWS architecture for their institutional platform, whilst ensuring service continuity and compliance.
[‘’]
AWS reste la référence cloud-native, avec un catalogue dont la profondeur n'a pas d'équivalent. C'est aussi le cloud qui demande le plus de discipline pour ne pas dériver côté facturation. Notre rôle, c'est d'aider les équipes IT à exploiter cette richesse sans tomber dans le piège du vendor lock-in.
Avec Terraform et un usage discipliné des standards CNCF, la trajectoire reste ouverte. C'est ce qui fait la différence entre une plateforme AWS qui s'autofinance par l'optimisation et une plateforme AWS qui dérive silencieusement.
Adrien Bresson, Head of Cloud Infrastructure
The fact that AWS is a mature platform does not mean that discipline is no longer required. The richer the ecosystem, the greater the risk of unintended complexity.
