The Explorer is the query tab of AI Observability. The Overview tab shows a fixed dashboard. In the Explorer, you write your own queries on LLM and AI trace data and pick one of four views for the results.
Use this page to run a query, filter it, pick a view, and open a trace.
Open the Explorer
- In the side navigation, open the More menu and select AI Observability.
- In the tab bar at the top, select Explorer.
All roles (Admin, Editor, and Viewer) can use the Explorer.
Run a query
- Type a filter in the filter bar, for example
trace.total_tokens > 1000. - Select Run Query, or press Cmd+Enter (Ctrl+Enter on Windows and Linux).
Fields with the trace. prefix hold the totals of a trace, such as trace.total_tokens, trace.estimated_total_cost, and trace.llm_call_count. You can filter and aggregate on these fields, which the Trace Explorer does not have.
The filter bar suggests only keys and values from GenAI spans. After you add a filter, the next suggestions show only values that match your current filters. For the filter syntax and the aggregation options, see Query Builder.
While a query runs, the Run Query button changes to Cancel. If you select Cancel, the view is empty until you run the query again.
To run the query again at an interval, turn on auto-refresh in the time picker.
Quick filters
Use the quick filters panel on the left to filter by common fields without typing a query. The panel has seven fields:
deployment.environmentgen_ai.operation.namegen_ai.provider.namegen_ai.request.modelservice.namegen_ai.tool.namegen_ai.agent.name
Each field lists RELATED values first. These values match the filters that you already applied. ALL VALUES lists the rest. When you select a value, the query changes and the active view shows the new results. To hide the panel, use the control at the top of the panel. To show it again, select Show Filters in the toolbar.
Views
The toolbar has four views: Trace View, List View, Time Series, and Table. All views use the same query. The Explorer opens in Trace View.
For trace totals, the views use different spans:
- Trace View uses all the spans of each trace, also the spans outside the selected time range.
- List View, Time Series, and Table use only the spans of each trace inside the selected time range. For example, with the filter
trace.total_tokens > 100, SigNoz adds up the tokens of the spans of that trace in the time range.
Trace View
Trace View shows one row for each AI trace, with totals for the trace such as tokens, cost, and call counts. To open the full trace, select a row. To open it in a new tab, hold Cmd (Ctrl on Windows and Linux) and select the row.
The trace opens in Trace Details. To show only the LLM spans of the trace, select the LLM quick filter at the top of that page.
The columns come from the trace fields that SigNoz finds for your data. By default, the table shows each of these fields that exists, in this order:
trace_id, service.name, root_span_name, estimated_total_cost, trace_duration_nano, span_count, total_tokens, input_tokens, output_tokens, distinct_tool_count, llm_call_count, tool_call_count, start_time, end_time, error_count, input, output, max_llm_duration_nano
Use the two controls above the table:
- Order by: sort the traces by a trace field, in ascending or descending order. The default is
last_activity_time (desc). You can sort by fields that the table does not show as columns, such aslast_activity_time. - Options: open the Edit columns drawer to show, hide, or reorder columns.
You cannot hide or remove the trace_id column, but you can drag it to a different position. Your browser saves your column selection, so the selection applies only in that browser.
List View
List View shows one row for each span. The default columns are timestamp, service.name, name, duration_nano, http_method, and response_status_code. The newest spans are at the top.
To change the columns:
- To remove a column, open the column actions in its header and select Remove column.
- To reorder a column, drag its header.
- To resize a column, drag the edge of its header.
The timestamp column stays first. Your browser saves your column changes. To open the trace of a span, select its row.
Time Series
Time Series shows the query results as a chart over time. If every query aggregates the span duration, the Y-axis unit is milliseconds. Otherwise, the unit is a plain number. To change the unit, use the unit control of the chart. You can also export the chart data.
To see exact cost values, use the Table view.
Table
Table shows the aggregated results of the query in a grid. To download the results, select the download icon, select csv or jsonl as the format, and then select Export.
Common tasks
| Task | Steps |
|---|---|
| Find the most expensive traces | In Trace View, set Order by to estimated_total_cost (desc). |
| Find failed tool calls | In List View, run the filter gen_ai.tool.name EXISTS AND has_error = true. |
| Chart token usage by model | In Time Series, set the aggregation to sum(gen_ai.usage.input_tokens) and group by gen_ai.request.model. |
| Compare cost by model | In Table, set the aggregation to sum(signoz.gen_ai.usage.tokens.cost) and group by gen_ai.request.model. The Cost over time panel on the Overview dashboard charts the same query. |
Related
- AI Observability overview: the Overview dashboard and the attributes that it needs.
- Trace Details: inspect the spans of one trace.
- Attribute Mapping: map non-standard attributes to GenAI names.
- LLM Observability integrations: instrument OpenAI, Anthropic, LiteLLM, CrewAI, and other frameworks.
Get Help
If you need help with the steps in this topic, please reach out to us on SigNoz Community Slack. If you are a SigNoz Cloud user, please use in product chat support located at the bottom right corner of your SigNoz instance or contact us at cloud-support@signoz.io.
