Kubernetes FAQs

FAQs cover the most common questions about Finout’s Kubernetes integrations. If you need additional assistance, contact Finout at [e-mail].

Is Finout double-counting Kubernetes costs running on a specific cloud infrastructure?

No. Finout does not double-count Kubernetes costs.

Kubernetes costs in Finout are not additional charges. Finout classifies the relevant cloud billing rows that are associated with Kubernetes resources as "Kubernetes" cost center. The total cloud cost remains unchanged.

When you filter by a cloud provider (for example, AWS), Kubernetes-related costs automatically appear under that provider because they are billed under the same cloud provider.

Finout does not increase the total cost; it only classifies the relevant billing rows as Kubernetes spend.

To view the costs of Kubernetes running on a specific cloud provider infrastructure:

Note: This scenario applies to any cloud provider that supports Kubernetes enrichment in Finout (AWS, GCP, or Azure).

This example uses AWS, but you can use the same steps for other supported cloud providers.

1. Open the MegaBill.

2. Open Filters and select the K8s Origin dimension.

   
3. Choose the relevant cloud provider (amazon-cur, for example) and click Apply Filters. This filters Kubernetes costs that run specifically on the selected cloud provider’s infrastructure.

   

Can different Kubernetes clusters in the same cloud environment use different monitoring tools when integrated with Finout?

Yes. When integrating with Finout, you can utilize various monitoring tools—such as Datadog, Prometheus, or other supported solutions—across multiple Kubernetes clusters within the same cloud environment (e.g., AWS, Azure, GCP, or Oracle Cloud). This approach enables you to tailor your monitoring strategy to the specific operational and financial needs of each cluster. For instance, a customer-facing production cluster might utilize Datadog for advanced observability, while an internal development cluster could employ Prometheus as a cost-effective alternative.

To ensure a successful integration with Finout in this setup:

1)Verify that each cluster operates independently and does not share node resources.

2)Configure each monitoring tool correctly to enable accurate data collection and ingestion into Finout.

Which GKE labels from GCP billing data does Finout support, and why?

Finout supports exactly two native GKE labels from GCP billing data:

• goog-k8s-cluster-name

• goog-k8s-node-pool-name

These are the only GKE-specific billing labels that are preserved. All other labels that start with k8s- or goog-k8s- are dropped during ingestion for GCP Compute Engine cost rows related to Kubernetes usage.

This is because these labels are already available from Kubernetes metrics. Ingesting them from billing data would duplicate information, significantly increase label cardinality, and cause performance issues. Kubernetes metrics are the authoritative source.

Where can I find the supported GKE labels in Finout?

The supported labels (goog-k8s-cluster-name, goog-k8s-node-pool-name) appear under Kubernetes Labels dimension, not under GCP dimensions.