> ## Documentation Index
> Fetch the complete documentation index at: https://docs.befailproof.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Metrics

> Plot magnitude metrics such as latency, cost, and token use.

Metrics is for values with units. Evaluations is for scores and judgments.

<Tabs>
  <Tab title="Dashboard">
    1. Go to **Observe → Metrics**.
    2. Select **add series** and choose an agent, environment, metric, statistic, and curve.
    3. Add series to compare agents or environments; different units receive separate axes.
    4. Select a point to open matching evaluations and sessions, or select **share** to copy the exact view.

           <img src="https://mintcdn.com/exosphere/WgPwQzedeDNwJBTy/images/dashboard/metrics-series.png?fit=max&auto=format&n=WgPwQzedeDNwJBTy&q=85&s=51c4ec844bb910c32b589085a967e6b2" alt="The Metrics page comparing two cost-per-output-unit series with time controls, aggregation selectors, curves, and a share action." width="1295" height="646" data-path="images/dashboard/metrics-series.png" />
  </Tab>

  <Tab title="CLI">
    There is no dedicated magnitude-metrics command yet. Discover available keys and inspect sessions or evaluations:

    ```bash theme={null}
    fp list score_filters
    fp --json evals --since 24h
    fp --json sessions --since 24h
    ```

    Use `fp query` for custom metric aggregation.
  </Tab>
</Tabs>