> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vantage.sh/llms.txt
> Use this file to discover all available pages before exploring further.
# Usage-Based Reporting
> Track cloud infrastructure consumption with Vantage Usage-Based Reporting.
To get started with Usage-Based Reports, you can also watch a video demo on [Vantage University](/vantage_university_cost_reporting) .
Usage-Based Reporting shows infrastructure consumption in non-monetary metrics, such as virtual machine running hours, stored gigabytes of data, and read/write units consumed. Cost Reports provide **Cost**, **Usage**, and [**Count**](/cost_reports#count-reporting) Y-axis display types. Select **Usage** to analyze provider-reported consumption; select **Count** to track how many distinct grouped items contribute cost over time.
## Supported Providers
The following providers populate usage data in Vantage. Units are normalized in the console for consistent spelling and capitalization (for example, `hrs` and `Hours` both display as **Hours**).
| Provider | Usage units | Examples |
| - | - | - |
| Anthropic | Tokens | Input, output, and cache token counts |
| Anyscale | Credits | Anyscale credits consumed |
| AWS | Dynamic (provider-metered) | Hours, GB-Months, Tokens, Requests, Units |
| Azure | Dynamic (provider-metered) | Quantity with Azure unit of measure (e.g., `10 GB`) |
| Azure CSP | Dynamic (provider-metered) | Hours, GB-Months, and other CSP unit types |
| Baseten | Hours, Tokens | Compute hours for Dedicated Inference and Training; input, output, and cached input tokens for Model APIs |
| Cartesia | Credits, Minutes | Credits for model capabilities (e.g., Text-to-Speech, Pro Voice Cloning); minutes for Line agents |
| CircleCI | Credits | CircleCI credits consumed |
| ClickHouse Cloud | CHC | ClickHouse Credits |
| Cloudflare | Dynamic (provider-metered) | Requests, DNS queries, rows, operations, and other Cloudflare usage units |
| Cursor | Tokens, Requests | Token totals for model usage; request counts for non-token usage |
| Custom Providers | User-defined | Values from `ConsumedQuantity` and `ConsumedUnit` in your CSV upload |
| Databricks | Dynamic (provider-metered) | DBU, DSU, GB |
| Elastic | Dynamic (provider-metered) | Hours and other Elastic line-item units |
| ElevenLabs | Credits, Minutes, Seconds, Milliseconds, Duration | Voice generation and related product usage |
| Fireworks AI | Tokens, GPU-Hours | LLM input/output/cached/training tokens; on-demand deployment GPU-hours |
| GitHub | Dynamic (provider-metered) | Quantity with GitHub `unit_type` |
| Google Cloud | Dynamic (provider-metered) | Hours, GB-Months, Tokens (Vertex AI), Requests |
| Grafana Cloud | Dynamic (provider-metered) | GB and other Grafana usage units |
| Kubernetes | vCPU-Hours, GB-Hours, GPU-Hours, GB | Allocated compute, memory, GPU, and network usage |
| OpenAI | Tokens | Token counts when usage enrichment data is available |
| OpenRouter | Tokens | Input, cached input, and output token counts by model and API key |
| Oracle Cloud | Dynamic (provider-metered) | GB-Months and other billed quantity units |
| Redis | Dynamic (provider-metered) | Hours and other FOCUS pricing units |
| SpaceXAI | Tokens, Images, Videos, Searches | Prompt, completion, cached, and reasoning text tokens; generated images and videos; web and X searches; code executions; input audio seconds; file storage |
| Temporal Cloud | Number, Gigabyte-hours, Pricing units | Actions, storage, and FOCUS pricing quantities |
| Twilio | Dynamic (provider-metered) | Numbers, minutes, messages, and other Twilio usage categories |
| Vercel | Dynamic (provider-metered) | FOCUS consumed or pricing units |
If a provider is not listed above, the **Usage Unit** column is empty and **Accrued Usage** displays 0.00. Vantage plans to add support for additional providers. Contact [support@vantage.sh](mailto:support@vantage.sh) with providers you'd like to see added.
Per-request LLM cost attribution by team, user, application, and feature is available when [Custom LLM Enrichment](/custom_llm_enrichment) is enabled and you deliver token telemetry to a connected S3 bucket. On Amazon Bedrock, the same attribution is available through [AWS Bedrock LLM Enrichment](/aws_bedrock_enrichment) once Model Invocation Logs are connected.
## Create Usage-Based Reports
Navigate to any Cost Report, or create a new report. By default, the **Y-Axis** displays **Cost**.
Apply any desired filters and **Group By** criteria. If the selected provider has multiple units of measure, use the **Usage Unit** filter to select the unit you want to analyze, such as **Hours**, **GB-Months**, or **Requests**. This filter is also available for Custom Providers and the **All Providers** filter context when the selected context contains multiple units.
Then, open the **Chart Settings** menu and select **Usage** for the **Y-Axis**.
Filter to a single Usage Unit so the report compares like-for-like consumption instead of combining incompatible units.
The following elements are displayed on the **Usage** view:
* The graph and chart display usage over the selected timeframe.
* The total cost is displayed as a trendline on the chart.
* **Usage Unit** is automatically selected and locked in the **Group By** menu. This prevents values with different units from being combined. If you switch the **Y-Axis** back to **Cost**, the grouping becomes optional and can be removed.
* In the table below the graph, the **Usage Unit** (e.g., GB-Hours or Metrics) and **Accrued Usage (a numerical value of calculated usage for that unit)** columns are displayed to show accrued usage for each grouping. The **Usage Unit** column is normalized for consistency across usage type for spelling, abbreviation, and capitalization.
* **Accrued Usage**, **Previous Period Usage**, **Accrued Cost**, **Previous Period Cost**, and **Change** between these cost values are also displayed. Click the **By Date** option above the table to see the usage data split out by date.
Save the report.
Usage data refreshes daily in the Vantage console. At this time, you cannot view retroactive usage metrics. Contact [support@vantage.sh](mailto:support@vantage.sh) if you want to view historical usage data.
### Export Usage Reports
To export usage reports, [follow the steps](/cost_reports#export-cost-reports) provided in the Cost Reports documentation. Ensure you are on the **Usage** view. **Usage Unit** is automatically selected as an export grouping so each unit of measure is identified in the exported data.
### View Usage Data for Custom Costs
When you create a [Custom Provider](/connecting_custom_providers), you can add usage data to your CSV upload files. Ensure you complete the `ConsumedQuantity` and `ConsumedUnit` columns. See the [Custom Providers documentation](/connecting_custom_providers#data-format) for information on the required schema.
## Understand Usage Units
The Usage Units displayed on Usage-Based Reporting include any dimension that a supported provider meters and bills. Vantage standardizes these dimensions for inconsistencies in unit case and spelling across providers. For example, Vantage standardizes `Hours`, `hrs`, `Hrs`, and `hours` to `Hours`. See the [Supported Providers](#supported-providers) table for the units available per provider.
When the selected provider context contains more than one distinct, non-empty unit, **Usage Unit** is available as a [Cost Report filter](/cost_reports#filter-by-usage-unit). If the context contains zero or one unit, Vantage omits the filter because it would not further narrow the report.
## Use Virtual Tags with Usage-Based Reporting
[Virtual Tags](/tagging) work across both cost and usage consumption for supported providers. If you created Virtual Tags based on cost, the associated usage consumption will inherit the same Virtual Tag. You can use Virtual Tags within your filter criteria to apply Virtual Tags based on usage.
## Usage-Based Reporting on Dashboards
You can add Usage-Based Reports to dashboards. Ensure **Usage** is selected for the **Y-Axis**, save the report, and then add it to a dashboard. Follow the [Dashboards](/dashboards) documentation for details on how to add a report to a dashboard.
## Usage Reports Examples
The following examples provide some common use cases for working with Usage-Based Reporting.
### Understand EC2 Running Hours
As an Engineering Manager, you want to view your Amazon EC2 usage by instance hours, so that you can get a profile of your compute usage. This is important because it helps you identify patterns in compute resource consumption. This type of report can help you create more efficient resource planning and cost optimization based on your team's workload needs.
You create the following Usage-Based Report in Vantage:
* Filter: All costs from **AWS** where Category **Amazon Elastic Compute Cloud - Compute** is **Compute Instance**
* Filter: **Usage Unit** is **Hours**
* Group By: **Subcategory**
The table below the chart includes the accrued usage and the unit of measure for each line item being viewed, such as EC2 instance bill code. To get the most value from this report, focus on identifying high-usage instance families and regions to optimize compute resources and explore opportunities for cost-saving strategies, like instance rightsizing or transitioning to Savings Plans.
### View S3 Storage by Tier
As a Storage Admin or Engineering Manager, you want to view detailed storage metrics across specific AWS services, such as the total data stored in S3 by tier. With this report, you can better understand the overall profile and efficiency of stored assets and potentially determine whether additional storage policies are required to optimize storage usage and growth.
You create the following Usage-Based report to see total data stored in S3 in Vantage:
* Filter: All costs from **AWS** where Category **Amazon Simple Storage Service** is **Storage**
* Filter: **Usage Unit** is **GB-Months**
* Group By: **Subcategory**
In the table below the graph, usage data is displayed in GB-Months for subcategories like:
* **TimedStorage-INT-AIA-ByteHrs** (Infrequent Access Tier)
* **TimedStorage-INT-IA-ByteHrs** (Intelligent Tiering)
* **TimedStorage-ByteHrs** (Standard Tier)
These metrics provide an overview of how data is distributed across different S3 storage tiers. Next steps could include analyzing whether data in the **Infrequent Access** and **Intelligent Tiering** categories is optimally stored, or you could decide whether more aggressive lifecycle policies are needed to move infrequently accessed data to cheaper storage tiers, like Glacier. You could also use this usage information to identify potential cost savings by reducing the storage volume in higher-cost tiers, such as the **Standard Tier**.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.