Cost Attribution in Shared Infrastructure

Short summary

A practical framework for attributing shared Kubernetes cluster costs across teams using a four-level maturity model: no attribution, tagging, show-back, and charge-back. It covers handling shared services, burst costs, and untagged resources with concrete recommendations. After three months of show-back, most teams cut 15-25% of waste through voluntary optimization driven by visibility.

• •Four-level cost attribution model: no attribution → tagging → show-back → charge-back
• •Shared services above 5% of total cost should be usage-based; smaller ones can be split equally
• •Untagged resources should default to platform team ownership to incentivize tagging enforcement
• •Three months of show-back typically yields 15-25% waste reduction before charge-back is needed

Generated with AI, which can make mistakes.