Platform reliability
Amazon EKS reliability & production-readiness review
For teams running EKS who aren't sure their cluster would survive a bad day.
What I check
- Node and autoscaling setup (Karpenter, Cluster Autoscaler, capacity)
- Workload resilience: probes, disruption budgets, requests and limits
- Security: IAM for pods, RBAC, network policies, image sources
- Observability, alerting and SLOs
- Upgrade path, backups and disaster-recovery readiness
From £950 typically 3–5 days
Includes: one EKS cluster in one AWS account, with a sample of up to three representative workloads.
You get
- A findings report with risks ranked by impact
- A prioritised fix list
- A one-hour walkthrough call
I run a four-cluster EKS platform at 99.5% availability, with an HA/DR design proven by DR exercises and chaos testing.