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

# Logs: view sessions, traces, and spans

> Find the traces your app sends to BotGauge. Browse sessions, traces, and spans in Logs, read the columns, and open a session for messages and timeline.

Traces you send with the [SDK](https://www.botgauge.com/docs/tracing/overview.md) or an [OTLP exporter](https://www.botgauge.com/docs/integrations/otlp.md) appear under **Observability, Logs**. Logs shows what each run of your agent did, with cost, duration, and evaluation scores.

## Sessions, traces, and spans

Logs has three tabs, each with a count.

| Tab | What it lists |
| - | - |
| Sessions | One row for each session, the conversation or run you identified with `session_id` |
| Traces | The traces recorded within sessions |
| Spans | The individual operations, such as LLM calls and tool calls |

Use the time filter (for example **All time**), the search box, and the **Column** menu to change what you see. The Column menu shows how many columns are visible, for example 8 of 8.

## The Sessions table

| Column | Meaning |
| - | - |
| Created | When the session started |
| Status | **Active** while the session is still receiving spans, otherwise **Inactive** |
| Session ID | The `session_id` you passed to `tracing.turn()`, with the agent name you passed to `tracing.agent()` shown beneath it |
| Traces | How many traces the session has |
| Duration | How long the session took |
| Cost | The cost, computed from the token counts on its LLM calls |
| Evaluation columns | One column for each judge that is turned on, for example **Answer Relevance** and **Correctness**. Values include Pass, Fail, Not evaluable, and Pending. |

Session IDs and agent names come straight from your code. A run traced with `tracing.turn(session_id="ses_langchain")` and `tracing.agent("langchain-agent")` shows that session ID, with `langchain-agent` as its label. See [Manual tracing](https://www.botgauge.com/docs/tracing/manual-tracing.md).

## Open a session

Click a row to open the session in a side panel. The panel shows the session's latency and cost, and arrows to move between sessions, for example "Session 1 of 135". It has three views.

| View | What it shows |
| - | - |
| Messages | The conversation as chat bubbles, with a tree on the left of the session, its turns, and each LLM call with its token counts |
| Trace Graph | The execution as a graph |
| Timeline | The spans along a timeline |

## Score your logs

Choose **Manage Evals** to change which evaluators score your sessions. Evaluators are called judges. See [Built-in judges](https://www.botgauge.com/docs/evals/built-in-judges.md).

## If your trace doesn't appear

Check that the API key belongs to the project you are viewing, because a key only reaches its own project. Run `botgauge probe` to confirm ingest accepted a span, and see [Tracing troubleshooting](https://www.botgauge.com/docs/tracing/troubleshooting.md).

Open Logs in the consoleGo to [BotGauge](https://cloud.botgauge.com) to browse sessions, traces, and spans in Observability, Logs.

[Try BotGauge →](https://cloud.botgauge.com)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.