# Dashboard

The Dashboard provides a **real-time snapshot** of your project's health and performance. Use the top bar to select date ranges, switch environments, and apply filters. Hover over any chart to see detailed information for that specific point in time.

### Overview Metrics

The top section displays key indicators with percentage changes from the previous period:

<div data-with-frame="true"><figure><img src="https://708770081-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F1ICwJbq7EJdn5kBgXnQu%2Fuploads%2F1m9H3vEBod6AkNdQP0u2%2Fimage.png?alt=media&#x26;token=b82f3d0d-c79c-4f44-a19a-5f0a4a8436be" alt=""><figcaption></figcaption></figure></div>

| Metric                    | Description                                   |
| ------------------------- | --------------------------------------------- |
| **Traces**                | Total number of traces recorded               |
| **Sessions**              | Total number of sessions                      |
| **Observations**          | Total observations captured across all traces |
| **Avg. Latency**          | Average response time across all traces       |
| **Avg. Session Duration** | Average length of sessions                    |
| **Scores**                | Total evaluation scores collected             |

### Detail Panels

Below the overview metrics, two panels provide granular breakdowns of activity and cost. Each panel includes a time-series chart and a sortable report table that you can export for further analysis.

<div data-with-frame="true"><figure><img src="https://708770081-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F1ICwJbq7EJdn5kBgXnQu%2Fuploads%2FaqBTtZTWcrP7WTbl6vU2%2Fimage.png?alt=media&#x26;token=3603f6e1-8dda-46c3-a067-33afc52f582a" alt=""><figcaption></figcaption></figure></div>

* **Traces** shows activity by trace name, including cost, number of runs, average cost, and latency for each trace type.
* **Models** breaks down usage by model, displaying cost, token consumption, cost per run, and average latency.