Updated 25 October 2026 · Published by Zoorik
Kubernetes cost allocation splits a shared cluster’s cost across its pods. You then add up pod costs by namespace, label or team. A pod’s share follows its requests, or the larger of its requests and its use.
How AWS EKS split cost allocation data works
Split cost allocation data is an opt-in feature for Amazon Elastic Kubernetes Service (EKS). It adds each pod’s cost to your AWS Cost and Usage Report. Its Kubernetes label support for cost allocation turns up to 50 labels per pod into cost allocation tags, which you then activate.
In AWS’s example, four pods share a 16 GB m5.xlarge node. Each is charged for the larger of its memory request and use:
- Pod1: requests 4 GB, uses 3 GB, charged for 4 GB
- Pod2: requests 4 GB, uses 6 GB, charged for 6 GB
- Pod3 and Pod4: request and use 2 GB each
- Idle: 2 GB that no pod requests or uses
AWS splits the idle cost among the pods by share size. Azure Kubernetes Service (AKS) cost allocation comes from its cost analysis add-on, which shows idle cost as its own charge.
Why it matters for your bill
- A pod that requests more than it uses is charged for its full request. That shows which teams hold unused capacity.
- Allocation splits the bill but doesn’t reduce it.
- Coral rebalances CPU and memory requests on AKS and EKS to match real use, so the same pods fit on fewer nodes.
- The free assessment works out your own figure.
Read next: AKS cost analysis: what the add-on shows and what it doesn’t
Sources: AWS, “Example of split cost allocation data” (opens in a new tab); AWS, “Understanding split cost allocation data” (opens in a new tab); AWS, “Using Kubernetes labels for cost allocation in EKS” (opens in a new tab); Microsoft Learn, “Azure Kubernetes Service (AKS) cost analysis” (opens in a new tab).