Projets

CEA — Commissariat à l'Énergie Atomique

RKE2
VMware
Monitoring
Longhorn
CI/CD

Technical Lead at the CEA (French Alternative Energies and Atomic Energy Commission), supporting the team on infrastructure automation, Kubernetes orchestration, observability and CI/CD for scientific computing workloads.

High-performance computing infrastructure

Context

The CEA is one of France’s leading research organizations. I intervened as Technical Lead to modernize their infrastructure, deploy a Rancher RKE2 cluster on VMware vSphere, and establish a complete DevOps toolchain for scientific application teams.

Responsibilities

  • Deployed and administered Rancher RKE2 clusters on VMware vSphere for container orchestration
  • Set up monitoring stack: Prometheus, Grafana dashboards, Centreon for infrastructure monitoring
  • Configured distributed logging with Elasticsearch/Kibana (ELK stack) and Alertmanager for alerting
  • Deployed Longhorn as distributed storage solution, combined with VMware vSphere CSI for persistent volumes
  • Wrote Ansible playbooks for fully automated cluster installation and configuration
  • Implemented Kyverno policies for security and compliance enforcement
  • Set up CI/CD pipelines with GitLab for continuous integration, image building, scanning and artifact management
  • Provisioned Artifactory as a container registry and binary repository

Key Achievements

  • Reduced infrastructure provisioning time by 80% through Ansible automation
  • Established a single-pane-of-glass monitoring view covering 50+ nodes
  • Enabled scientific teams to self-serve container deployments with guardrails via Kyverno

Technical environment: VMware vSphere, Rancher RKE2, GitLab, Kyverno, Artifactory, Elasticsearch, Grafana, ArgoCD, Ansible, Longhorn, Prometheus, LDAP