Yotascale now supports the allocation of Kubernetes cost by namespace, pod, deployment, and label. Yotascale uses a combination of metrics gathered from your cluster and the actual cost of associated instances to show the actual cost of each grouping. This means that if the instances serving your Kubernetes infrastructure are receiving a Reserved Instance discount, that amount will be used in the namespace, pod, deployment, or label calculation.
Kubernetes cost is broken out as a subset of your EC2 costs - so you can see your total EC2 cost as it's shown in your AWS bill, and the subset of that EC2 cost that is related to your containerized infrastructure.
To get started, install the Yotascale agent...
Frequently Asked Questions
- Does Yotascale support ECS and EKS allocation?
EKS is supported with the use of the Yotascale Kubernetes agent.
ECS is supported by enabling Container Insights on your ECS clusters.
- Can I group Kubernetes and non-Kubernetes costs together?
Yes! Let us know how you'd like to group your resources and we can help. You can describe non-Kubernetes resources using tags, accounts, and other existing identification information, and the corresponding Kubernetes resources by namespace, deployment, or label.
An example grouping might be: any resource where Application = 'cost-analytics' OR Namespace = 'cost-analytics'
This allows us to break apart Kubernetes costs for organizations and allocate cost back to the teams who are benefitting from Kubernetes as a platform or organization-wide offering.
- How many versions of the agent are supported at a given time?
We currently support up to 3 minor versions of the agent at one time.
- How does Yotascale calculate the cost per pod?
Yotascale uses the CPU and RAM utilization metrics for each pod, the cost of the EC2 instance the pod was running on, and the duration the pod was active to calculate the cost of each pod. Using Namespace or Labels, we can map the pod back to the team that owns it and the cost of all the team's infrastructure can be viewed in one place.