Running 18 EKS clusters with a 2-person platform team

The Setup

At Nielsen Media, our small platform team owned 18 production EKS clusters supporting a diverse set of applications. The challenge: operate them consistently, ship quickly, and keep reliability high.

What Worked

• Modular Terraform for cluster provisioning so every new environment looked the same.
• Helm charts for the standard add-ons (Prometheus, Grafana, ingress, autoscaling).
• Argo CD for GitOps so the desired state of every cluster and application was auditable in git.
• DORA metrics to measure deployment frequency, lead time, and change failure rate.

What I’d Do Differently

• Invest in cluster templating earlier — we ended up rewriting several modules after the third or fourth cluster.
• Standardize on a single observability stack from day one to avoid cross-cluster query pain.

Key Takeaway

Small platform teams can operate many clusters when the platform is treated as a product: opinionated defaults, repeatable modules, and strong feedback loops.