Show HN: I audited 500 K8s pods. Java wastes ~48% RAM, Go ~18%
3 days ago
- #DevOps
- #Cost Optimization
- #Kubernetes
- Detects over-provisioned CPU and memory in Kubernetes clusters, typically by 30-60%.
- Analyzes pod resource configurations to identify gaps between set limits and actual needs.
- Estimates cost waste, with an example showing $14,880 annual waste detected.
- Breaks down waste by category: Memory, CPU, and orphaned Load Balancers.
- Identifies the biggest waster, e.g., 'api-gateway' pod wasting $8,400/year.
- Provides fixes like adjusting memory requests to match actual usage.
- Lists common issues: Over-provisioned memory/CPU, orphaned load balancers, unbound volumes.
- Uses read-only kubectl commands for analysis, requiring metrics-server for accurate usage data.
- Offers dashboard features for detailed analysis, historical trends, and PDF export.
- Operates locally with optional data push to dashboard; no agents installed.
- Uses conservative cloud pricing averages for cost estimation.
- Supports disabling telemetry and is open source, encouraging community contributions.