Source

This page is generated from skills/eks-cost-intelligence/README.md. Edit the source, not this page.

eks-cost-intelligence

A Kiro agent skill that connects to a live EKS cluster, analyzes cost signals across 6 dimensions, and produces a scored cost intelligence report with dollar-quantified findings and prioritized remediation.

What it does

Unlike static cost best-practices guidance (eks-best-practices), this skill performs live assessment against a specific cluster. It pulls data from three sources and correlates them to produce dollar-denominated waste findings:

• AWS Cost Explorer — actual spend by cluster, namespace, and workload
• CloudWatch Container Insights — real CPU/memory utilization (P50/P95)
• Kubernetes API — resource requests, limits, replica counts, PVCs, Services

The output is a Cost Score (0–100) with classification (OPTIMIZED/GOOD/FAIR/NEEDS_WORK/CRITICAL) and a prioritized findings table with dollar impact and ready-to-apply remediation snippets.

Assessment Dimensions

Dimension	Weight	What It Measures
Compute Efficiency	25 pts	CPU/memory waste, over-provisioning, missing requests
Spot/Graviton Adoption	20 pts	Spot %, Graviton eligibility, instance diversity
Networking Costs	15 pts	Cross-AZ traffic, VPC endpoints, topology routing
Storage Costs	15 pts	gp2→gp3, unused PVCs, oversized volumes
Observability Costs	10 pts	Control plane logging, metric cardinality
Idle Resources	15 pts	Zero-scale deploys, orphaned LBs, empty namespaces

Structure

skills/eks-cost-intelligence/
├── SKILL.md                              # Skill entry point (frontmatter + workflow)
├── README.md                             # This file (contributor documentation)
├── references/
│   ├── compute-efficiency.md             # Dimension 1 checks
│   ├── spot-graviton-adoption.md         # Dimension 2 checks
│   ├── networking-costs.md               # Dimension 3 checks
│   ├── storage-costs.md                  # Dimension 4 checks
│   ├── observability-costs.md            # Dimension 5 checks
│   ├── idle-resources.md                 # Dimension 6 checks
│   ├── fargate-costs.md                  # Fargate detection + Fargate-specific checks
│   ├── report-generation.md              # Scoring algorithm + report template
│   ├── cost-data-collection.md           # API calls for data sources
│   ├── waste-calculation.md              # Dollar waste formulas
│   ├── cost-estimation-fallback.md       # Node-based estimation fallback
│   └── findings-format.md               # Output schema + remediation templates
└── tools/
    └── report_to_html.py                 # Markdown → HTML converter (stdlib only)

Prerequisites

• AWS credentials with EKS read access
• kubectl configured for the target cluster
• Required permissions: eks:DescribeCluster, eks:ListClusters, eks:ListNodegroups, ec2:DescribeInstances, ec2:DescribeVolumes, elasticloadbalancing:DescribeLoadBalancers
• Optional (for richer analysis): ce:GetCostAndUsage, cloudwatch:GetMetricData, pricing:GetProducts, elasticfilesystem:DescribeFileSystems, elasticfilesystem:DescribeLifecycleConfiguration

Relationship to other skills

Skill	Relationship
eks-best-practices	Provides advisory guidance; this skill adds the live dollar layer
eks-operation-review	Assesses operational health; this skill focuses on cost
eks-recon	Discovers cluster state; useful to run before cost analysis
eks-upgrade-check	Assesses upgrade readiness; complementary but different concern

Pricing Data

This skill uses dynamic pricing lookups (AWS Price List API) as the primary method for dollar estimates. A static reference pricing table is included as a fallback for environments where the Price List API is unavailable. The static table was last verified in June 2026 and covers us-east-1 On-Demand rates.

For production assessments, ensure pricing:GetProducts permission is available for the most accurate results.

Multi-Cluster Support

Multi-cluster fleet-wide assessment is out of scope for the initial release. The skill assesses one cluster at a time. For fleet-wide cost analysis, run the assessment against each cluster individually and aggregate the reports externally.

Future enhancement: a multi-cluster.md reference may be added to support fleet-level aggregation and cross-cluster comparison.

Contributing

This skill is developed in-repo (not vendored from an upstream source). To contribute:

1. Follow the patterns established by sibling skills (eks-operation-review, eks-upgrade-check)
2. Reference files use progressive disclosure — loaded on-demand per dimension
3. All pricing data should prefer dynamic API lookups over static tables
4. Remediation snippets should be ready-to-apply (kubectl, YAML, Terraform)
5. Run the eval scaffold (misc/evals/eks-cost-intelligence/) to validate changes

See the repository's CONTRIBUTING.md for general guidelines.

---

This skill is provided as sample code for educational and demonstration purposes only.