> For the complete documentation index, see [llms.txt](https://docs.vngcloud.vn/vng-cloud-document/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.vngcloud.vn/vng-cloud-document/ai-stack/agent-base/usage-budget/view-usage-cost.md).

# View Usage & Cost

> Guide for Root, Admin, and Viewer to monitor resource consumption (MaaS and Runtime/Agent), cost breakdown, and month-end cost projection on the **Usage & Cost** dashboard.

***

## Prerequisites

* Signed in with role **Root**, **Admin**, or **Viewer** to see full org data
* Or role **Member** to see data for agents you created

***

## Open Usage & Cost

1. Go to **AgentBase** → sidebar → **Usage & Budget** → **Usage & Cost**
2. The dashboard opens on the **Usage tab** by default with Time Range = **Last 1h**

***

## Apply Filters

The global filter bar is pinned at the top of the page and applies to both the Usage and Cost tabs:

| Filter           | Description                                                                                         |
| ---------------- | --------------------------------------------------------------------------------------------------- |
| **Agent**        | Multi-select — choose specific agents or **Others** (direct MaaS calls not routed through an agent) |
| **All Provider** | Multi-select — filter by LLM provider (OpenAI, Anthropic, VNG AIP, ...)                             |
| **All Models**   | Multi-select — filter by AI model (gpt-4o, claude-sonnet-4-6, ...)                                  |
| **All API Keys** | Single select — filter by API Key                                                                   |
| **Time Range**   | Quick bar + picker panel                                                                            |

***

## Select a Time Range

**Quick bar:** Click any preset directly: **15m · 30m · 1h · 2h · 3h · 6h · 12h · 1d**

**Picker panel** (click **∨** to open):

| Tab          | How to use                                                                                            |
| ------------ | ----------------------------------------------------------------------------------------------------- |
| **Quick**    | Select a preset: Last 15 minutes through Last 6 months; Today, This week, This month, Last month, ... |
| **Relative** | Enter a custom relative range (e.g., Last 45 minutes)                                                 |
| **Absolute** | Select a specific From date and To date; validates that To ≥ From                                     |

***

## Usage Tab — View Usage Metrics

The Usage tab shows 4 KPI cards and 3 main sections:

**KPI cards:**

| Card                | Description                                                             |
| ------------------- | ----------------------------------------------------------------------- |
| **Total Requests**  | Total API calls within the filter range                                 |
| **Tokens Consumed** | Total Input Tokens + Output Tokens (including Cache)                    |
| **Active Agents**   | Number of agents with at least 1 request — Others (MaaS) is not counted |
| **Errors**          | Total number of failed requests within the filter range                 |

**Charts:**

* **Requests over time**: Line chart — X-axis grouped by hour (≤ 1 day) or by day (> 1 day); hover tooltip shows breakdown by model and total request count
* **Token Breakdown**: Doughnut chart separating Cache / Output / Input tokens — hover to see values and percentages

**Top Expensive Agent table:**

Shows the top 10 agents by total cost within the selected time range:

| Column          | Description                                         |
| --------------- | --------------------------------------------------- |
| **Agent**       | Agent name with icon                                |
| **Tokens Used** | Total tokens consumed                               |
| **Cost (VND)**  | Total cost (VND), sorted by highest cost by default |
| **Started At**  | Timestamp of the most recent session start          |

An **Others** row appears at the bottom of the table when there is MaaS usage not attached to any agent.

**Budget warning banner:**

When current month spending reaches ≥ 80% of the budget limit, a warning banner is pinned to the top of the Usage tab:

* **80–99%**: Yellow banner — "⚠️ 80% of this month's budget used. Consider reducing load or reviewing active agents." \[**View budget** link]
* **100%+**: Red banner — "⚠️ Monthly budget exceeded. Immediate action required."

***

## Cost Tab — View Cost Breakdown

Switch to the **Cost tab** for detailed cost analysis:

**KPI cards:**

| Card                   | When shown                                   |
| ---------------------- | -------------------------------------------- |
| **Total Cost**         | Always visible                               |
| **Budget Used**        | Only when Root has set a budget limit        |
| **Monthly Projection** | Always visible (after ≥ 3 days in the month) |

**Cost breakdown:**

* **Cost by Agent**: Table + chart; click an agent row → filters the entire dashboard to that agent
* **Cost by Provider**: Doughnut chart — OpenAI / Anthropic / VNG AIP with % contribution
* **Cost by Model**: Table + chart — Model Name, Provider, Total Cost, % of Total; sortable by cost descending
* **Monthly Cost Trend**: Bar/line chart — daily cost for the current month; hover tooltip shows date, total cost, and request count

***

## Role-based Visibility

| Role       | Data visible                                                                            |
| ---------- | --------------------------------------------------------------------------------------- |
| **Root**   | Full org data (all agents, providers, models, MaaS/Runtime) + access to Budget & Alerts |
| **Admin**  | Full org data — no access to Budget & Alerts                                            |
| **Viewer** | Full org data — no access to Budget & Alerts                                            |
| **Member** | Only data for agents they created; **Others** (MaaS) option is not shown                |

***

## Result

| I want to...                          | Go to                                                                                             |
| ------------------------------------- | ------------------------------------------------------------------------------------------------- |
| Set a budget limit and receive alerts | [Configure Budget & Alerts](/vng-cloud-document/ai-stack/agent-base/usage-budget/budget-alert.md) |
| Usage & Budget overview               | [Usage & Budget](/vng-cloud-document/ai-stack/agent-base/usage-budget.md)                         |


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.vngcloud.vn/vng-cloud-document/ai-stack/agent-base/usage-budget/view-usage-cost.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.