The challenge
This global pharmaceutical leader runs one of the world’s largest pharmaceutical operations on a Kubernetes footprint that spans AWS and Azure, EKS and OpenShift. The estate was significantly overprovisioned.
There was no automated way to rightsize it. The team needed waste identification, observability, rightsizing, and automation in one solution, not four. At this scale, manual tuning was never going to keep up.
The solution
The customer deployed the Kubex Kubernetes Automation Controller in its development environment.
Kubex identified the waste, calculated the right CPU and memory levels based on actual workload behavior, and executed the rightsizing automatically. The platform team did not have to choose between trusting automation and keeping control. Kubex gave them both inside the guardrails they defined.
The results
- $585K in annualized savings
- 907 CPU cores reclaimed
- 8.8TB of memory reclaimed
- Delivered in two weeks of automated rightsizing
- Foundation in place to extend optimization to GPU workloads next
In their words
“The savings has been a really powerful story to tell our leaders about the outcomes of our collaboration.”
FinOps and cloud engineering leadership, global pharmaceutical leader
