Case Study: NOS cuts Kubernetes costs by over 50% with PerfectScale

NOS cuts Kubernetes costs by 50% with PerfectScale

NOS, one of Portugal's leading telecom providers, faced significant challenges in optimizing its complex hybrid Kubernetes environment. Using tools like Prometheus and Grafana provided raw metrics but no actionable recommendations, leading to failed manual optimization attempts that caused crashes and SLA breaches. This broke internal trust and forced teams to overprovision resources to avoid risk, while FinOps spent days each month piecing together fragmented cost data. They partnered with PerfectScale to find a safer way to cut costs.

The solution from PerfectScale provided smart automation and contextual evidence for safe optimization. After a successful rollout in development, NOS expanded its use to production, where it reduced costs by over 50% on their largest cluster and eliminated idle resources on several main node pools. PerfectScale also uncovered and resolved performance issues like out-of-memory kills. The results included improved application performance, reclaimed FinOps time, and rebuilt company-wide trust in optimization, enabling adoption across engineering and finance teams.

NOS