Today, Vantage announces the general availability of Anyscale cost support, allowing Vantage customers to track their costs and usage of Anyscale alongside their other infrastructure providers. Customers running AI and ML workloads on Anyscale can add any number of accounts from the Integrations page in order to visualize costs and usage, as well as utilize the Vantage Kubernetes Agent to monitor workload efficiencies.

Anyscale is a cloud platform built on the Ray framework that makes it easy to scale and deploy Python applications and AI/ML workloads without managing complex distributed systems. Anyscale supports deployment via a public cloud provider like AWS or GCP, on-prem with Machine Pools, or into your existing Kubernetes cluster. For a complete view of application costs, teams need to see their Anyscale and other cloud provider expenses in a single Vantage report. This was previously possible through the Custom Provider integration in Vantage, but this integration required extra manual work to retrieve, format, and upload the data. Customers could also deploy the Vantage Kubernetes agent for utilization monitoring, but lacked cost data to attribute costs.

Now, with the launch of Anyscale cost support, customers can grant Vantage access to their Anyscale cost data via an API key, and Vantage will ingest and report on daily usage data. This data can be grouped and filtered by dimensions such as organization, project, cloud provider, instance type, and more. Customers can allocate usage across teams through Virtual Tags, detect anomalies in workload scaling, and compare Anyscale consumption alongside AWS, GCP, and other providers to get a full view of application costs. Customers can enable even more granular reporting by installing the Vantage Kubernetes Agent on Anyscale clusters in order to monitor GPU efficiency and allocate costs by dimensions such as label or namespace.

The Anyscale integration is now available to all Vantage customers. To connect, go to the Integrations page in account settings and select Anyscale. You can learn more via the new Anyscale section in Vantage’s documentation.

Frequently Asked Questions

1. What is being launched today?

Vantage is launching native support for Anyscale cost and usage data. Customers can now connect Anyscale via an API key, enabling automatic ingestion and visibility of Anyscale usage and costs in Vantage.

2. Who is the customer?

Any Vantage customer that uses Anyscale to run workloads and wants visibility into instance usage, allocation, and optimization.

3. How much does this cost?

There is no additional fee for using the integration. However, Anyscale costs will be included in quota tier enforcement. In the event that your Anyscale costs push you over your current tier limit, you may be prompted to upgrade. To see more details on pricing, please refer to the Pricing page.

4. How does the Anyscale integration in Vantage work?

Customers create an API credential in the Anyscale API key generation console and connect it to Vantage via the Integrations page (see detailed steps in our documentation). Vantage then ingests cost and usage data using the Aggregated Instance Usage API. This data is normalized and displayed in reports, where it can be filtered and grouped by organization, project, cloud provider, and more.

5. What permissions are required on the Anyscale side?

You must be an Organization Owner to access the Aggregated Instance Usage API.

6. What permissions are needed within Vantage?

You must have a Vantage Owner or Integration Owner role to add or remove the Anyscale integration. For more information, see Vantage’s Role Based Access Control documentation.

7. What permissions does Vantage have in my Anyscale account?

Anyscale only has a single type of API key that can be used to programmatically access Anyscale APIs. This key includes additional permissions beyond what Vantage uses. Vantage only reads data from the Aggregated Instance Usage endpoint and will never perform any other actions.

All API Keys generated through Anyscale have revocable read-write access. Unfortunately, Anyscale does not currently support more fine-grained permissions for Admin Keys. We recommend customers submit a feature request to Anyscale for improved permission scoping, and Vantage will adopt fine-grained permissions if this becomes available.

8. What dimensions can Anyscale costs be filtered and grouped by?

Anyscale reports can support aggregating and filtering on the following dimensions:

• Billing Account (Organization)
• Account (Project)
• Service (e.g., Instances)
• Category (e.g., customer_hosted)
• Subcategory (e.g., k8s)
• Resource ID (Resource ID)
• Charge Type (e.g., Usage)
• Cloud Provider (Available as a tag anyscale:cloud_id or anyscale:cloud_provider)
• Instance Type (Available as a tag anyscale:instance type)
• User (Available as a tag anyscale:user_id or anyscale:user_email)
• Workspace (Available as a tag anyscale:workspace_id)

9. Can I view usage data for Anyscale?

Yes, usage data is available for Anyscale consumption, measured in Credits.

10. Are there Active Resources for Anyscale?

No, there are no Active Resources to monitor for Anyscale.

11. How often does Anyscale data refresh in the Vantage console?

Anyscale data refreshes daily in the Vantage console.

12. What happens if I remove an Anyscale integration?

If you decide to remove your Anyscale integration from Vantage, all associated costs will be removed from the Vantage console associated with that Anyscale API Key.

13. How far back will the data go?

The integration will fetch historical data up to your account’s configured retention period.