How a Saudi Development Bank Revived Dormant Cloud Infrastructure
4 Kubernetes clusters revived, 10+ applications deployed, $180K annual savings from idle resources.
The challenge
A major development bank in Saudi Arabia had invested significantly in cloud infrastructure over several years, but organizational changes and vendor transitions left the environment in a state of disrepair. Four Kubernetes clusters sat idle, dozens of partially configured services consumed budget without serving traffic, and the internal team lacked the institutional knowledge to safely resume operations.
Previous vendors had left minimal documentation — a common problem given HashiCorp's finding that 74 percent of organizations report platform engineering and security teams do not work as a unified function. The bank was paying over $15,000 per month for infrastructure that served no production workloads, and leadership needed a clear assessment: revive the existing investment or start from scratch.
The approach
Nepheli conducted a full infrastructure audit using our knowledge-graph methodology, mapping every resource, dependency, and configuration state across the bank's AWS environment. The audit classified resources into three categories: recoverable, rebuild-required, and decommission.
The audit itself identified $180,000 per year in idle resource costs that could be eliminated immediately — orphaned load balancers, unattached storage volumes, running instances with no network connectivity, and expired reserved instances still being renewed. This pattern is typical: DataStackHub's research found that idle resources account for 10 to 15 percent of monthly cloud invoices and orphaned storage adds another 3 to 6 percent. These savings funded the recovery engagement.
The execution
Over a 12-week engagement, we brought all four Kubernetes clusters back to production readiness. This involved updating control plane versions, rebuilding node groups with current AMIs, migrating from deprecated APIs, and implementing proper namespace isolation and RBAC policies that had been absent in the original deployment. With 96 percent of enterprises now running Kubernetes according to the CNCF, reviving existing clusters was the right call over starting from scratch.
We deployed over 10 internal applications onto the revived clusters, implementing CI/CD pipelines with automated compliance checks aligned to the bank's regulatory requirements. Each application received a standardized deployment template, monitoring configuration, and incident response runbook.
The results
The bank's cloud infrastructure went from a dormant cost center to an active platform serving internal applications within 12 weeks. Annual cloud spend decreased by $180,000 from idle resource elimination alone — consistent with Flexera's finding that organizations waste 27 to 32 percent of cloud budgets — while the platform now supports over 10 production workloads that previously ran on on-premises hardware.
The bank's platform team was trained on the new architecture and is now operating independently with full documentation and runbook coverage. The engagement established a model for the bank's ongoing cloud strategy, with clear governance policies and operational procedures that prevent the documentation and knowledge gaps that led to the original problem. The cost optimization methodology used in this engagement has since become a standard part of Nepheli's infrastructure assessment process.