Case study
Using Kyverno scale-down policies managed through Nirmata, the platform team scaled down 1,945 idle deployments across two clusters and released 579 CPU cores and 4.5 TiB of memory.
Challenge
The company runs large, shared Kubernetes clusters serving hundreds of application teams. Over time, deployments that were no longer in active use kept holding their CPU and memory requests. That reserved capacity kept nodes running and made clusters look full, even though much of it was doing no work.
Solution
The platform team used Kyverno, the CNCF policy engine, with Nirmata to enforce scale-down as policy. Rules apply the same way across every namespace, so no team has to chase owners or run manual clean-up projects, and every action is recorded in a policy report.
Scaled-down workloads keep their configuration, so owners can restore them in seconds. Nothing is deleted without the allow-delete marker.
Results
Policy reports dated 23 September 2026. Monthly and annual cost use the standard pricing described below.
| Cluster | Deployments scaled down | Marked allow-delete | Namespaces affected |
|---|---|---|---|
| Cluster 1 | 1,121 | 14 | 393 |
| Cluster 2 | 824 | 7 | 256 |
| Total | 1,945 | 21 | 649 |
| Cluster | CPU (cores) | Memory (GiB) | Monthly | Annual |
|---|---|---|---|---|
| Cluster 1 | 301.7 | 2,931.8 | $16,030 | $192,361 |
| Cluster 2 | 277.4 | 1,607.7 | $11,374 | $136,487 |
| Total | 579.1 | 4,539.5 | $27,404 | $328,849 |
| Monthly cost released | CPU | Memory | Total |
|---|---|---|---|
| Cluster 1 | $6,962 | $9,068 | $16,030 |
| Cluster 2 | $6,401 | $4,973 | $11,374 |
| Total | $13,363 (49%) | $14,041 (51%) | $27,404 |
Memory accounts for slightly more than half of the released cost. Cluster 1’s workloads were especially memory-heavy, at about 9.7 GiB per released core.
Next steps
The team plans to roll the same policies out to the rest of the fleet, turn the 21 allow-delete candidates into permanent clean-up, and pair scale-down with right-sizing so active workloads request only what they use.
Method
Case study
A leading quick-service restaurant chain streamlined its governance and compliance processes across multiple Amazon EKS clusters with Nirmata Control Hub, the Unified Governance Platform.
Case study
Chase UK made Kubernetes audit-ready by replacing a manual compliance process with Kyverno policies enforced in every cluster from the day it is built. Policy reports feed one Grafana dashboard for developers, security teams and auditors.
Case study
Razorpay meets the Reserve Bank of India’s Payment Aggregator Master Directions across more than 7,000 Kubernetes nodes by writing its regulatory controls as Kyverno policies. The same policies run in CI, at admission and in periodic scans, so every change is checked against RBI requirements and auditors see live enforcement records instead of spreadsheets.
