Container Service for Kubernetes:Custom cost accounting policies

When to use custom pricing

Custom pricing templates are useful in the following scenarios:

• An internal operations team purchases compute resources in bulk and resells them to business units at different rates.

• Pods with identical specifications land on nodes of different specifications due to scheduling. You want to standardize costs at the business level and eliminate the resulting cost variation.

• Business units deploy Pods with the same resource request but different overcommit ratios on the same node, and you need differentiated cost accounting based on the resource limit.

Prerequisites

Before you begin, make sure you have:

• Enabled cost insights

• Obtained the kubeconfig file for your cluster and connected to the cluster using kubectl

Step 1: Enable custom pricing in cost-exporter

1. Log on to the ACK console. In the left navigation pane, click Clusters.

2. On the Clusters page, click the name of your cluster. In the left navigation pane, click Workloads > Deployments.

3. At the top of the page, select kube-system from the Namespace drop-down list. Find the ack-cost-exporter Deployment. In the Actions column, choose More > View in YAML. In the command section, add '--enable-custom-price=true'. Click Update.

   

Step 2: Configure a pricing template

1. Create a file named price-template.yaml with the following content.

   Important

   All fields under spec are optional. A Pod matches a template when all configured spec fields match the Pod's metadata. A template with an empty spec matches all Pods. If a spec field is omitted, that condition is not applied and the template matches Pods regardless of that attribute.

   Templates are evaluated in order. When a Pod matches multiple templates, the system uses the first matching template from top to bottom.

   apiVersion: v1
   kind: ConfigMap
   metadata:
     name: price-template
     namespace: kube-system
   data:
     template.yaml: |
       - name: template-nginx
         spec:
           labelSelector:
             app: nginx
           namespaceSelector:
             - default
             - kube-system
           resource:
             cpu: 500m
             memory: 512Mi
             cpu-overcommit-ratio: 2
             memory-overcommit-ratio: 2
         price:
           cpu: 0.2
           memory: 0.03
       - name: template-default
         price:
           cpu: 0.1
           memory: 0.02

   This template defines two pricing rules:

   • template-nginx applies to Pods that meet all of the following conditions: the Pod has the app: nginx label, runs in the default or kube-system namespace, has a CPU resource request of 500m (limit: 1 Core), and has a memory resource request of 512 MiB (limit: 1 GiB).

   • template-default applies to all other Pods in the cluster.

   The following table describes the template parameters.

   Parameter	Description
   name	Name of the pricing template.
   spec.labelSelector	Selector for matching Pod labels. Omit to match Pods with any label.
   spec.namespaceSelector	List of namespaces to match. Omit to match Pods in any namespace.
   spec.resource.cpu	CPU resource request of the Pod. Omit to match Pods with any CPU request.
   spec.resource.memory	Memory resource request of the Pod. Omit to match Pods with any memory request.
   spec.resource.cpu-overcommit-ratio	CPU overcommit ratio of the Pod, calculated as CPU resource limit / CPU resource request. Omit to match Pods with any CPU overcommit ratio.
   spec.resource.memory-overcommit-ratio	Memory overcommit ratio of the Pod, calculated as memory resource limit / memory resource request. Omit to match Pods with any memory overcommit ratio.
   price.cpu	Unit price of the Pod's CPU. Unit: Core-Hour.
   price.memory	Unit price of the Pod's memory. Unit: GiB-Hour.

2. Apply the ConfigMap to your cluster.

   kubectl apply -f price-template.yaml -n kube-system

What's next

• Estimate Pod costs based on cluster scheduling watermarks using single-resource (CPU or memory) or weighted-mix estimation. See Introduction to cost estimation policies.

• Query cost insights data through the HTTP API for custom development. See Overview of calling an API to query cost data.