Skip to content

Explorer

After the application is created successfully, the system aggregates each Agent/LLM request as a log event into the Explorer. You can search, filter, and analyze all Session and Trace data of a single or all applications.

The top of the page supports switching between Agent and LLM status. The default view is Agent status.

Agent Status

In Agent status, two filter boxes are displayed at the top of the page:

  • Agent Type: Filter by all Agent types or select a specific type.
  • Instance: Filter by a specific instance under the selected Agent type. Instance names are searchable. When a specific Agent type is selected, the instance filter only shows instances under that type. When no Agent type is selected, the instance filter defaults to "All Instances" but is not selectable.

LLM Status

When you switch to LLM status, an application filter appears at the top of the page. You can select all applications or a specific LLM application to view.

View Mode

User List

The Agent total Explorer supports aggregating data across Agent types by User dimension. The detail page of a single Agent type also provides a User tab for viewing user usage under that type.

The Agent type detail page includes the following first-level tabs:

Instance List / User List / Session / Trace

Session List

By default, Session-level data is displayed. The list fields include Session ID, Input Tokens, Output Tokens, Risk Events, and Alert Severity. The left-side quick filter supports filtering by token range, model provider, request model, response model, and other dimensions. Multiple analysis views are available: List, Top List, Time Series, Pie Chart, Treemap, and Grouped Table.

  • Risk Events: Total number of risk events detected in the current Session.
  • Alert Severity: The highest alert severity among the risk events in the current Session. If no risk events exist, it shows "No Risk."

Click a row to enter the Session detail page:

  • Overview: The left side shows the Session Trace waterfall chart; the right side shows the preview, property information, and the alert severity of the current Session.
  • Invocation Analysis: Displays the proportion of model calls, Tool executions, and Skill executions, along with the Traces and risk events associated with the current Session. Risk events show alert severity, associated operation, and occurrence time.

Trace List

Switch to the Trace List tab to view data by Trace dimension. The list fields include Trace ID, Associated Session, Trace Duration, Input Tokens, Output Tokens, Trace Status, Risky Operations, and Alert Severity.

  • Risky Operations: Number of operations marked as risky in the current Trace.
  • Alert Severity: The highest alert severity among all risky operations in the current Trace.

Click a row to enter the Trace detail page:

  • Overview: Displays the Trace invocation process using a waterfall chart, along with the alert severity of the current Trace. Below, Span details are shown.
  • Invocation Analysis: Displays Tool and Skill invocation records. Risky operations are marked with a risk indicator, and the risk level is shown along with the relevant Tool, Skill, and other operation information.

Detail Page

When you enter the detail page, the left side displays the individual Span and Trace generated by the application, while the right side shows the corresponding preview and property information.

Risk Alert

Based on risk events generated during Agent execution, the system applies risk labels to objects such as Session, Trace, Tool, and Skill.

Risk alert information may appear in the following locations:

Location Display Content
Session List Risk Events, Alert Severity
Session Overview Alert severity of the current Session
Session Invocation Analysis Risk events and associated Traces
Trace List Risky Operations, Alert Severity
Trace Overview Alert severity of the current Trace
Trace Details Risky operation indicators, and risk levels of operations such as Tool and Skill

When a Session or Trace contains multiple risk events, the list and overview display the highest alert severity. In the Invocation Analysis section, you can click a specific risk event or risky operation to navigate to the Security Check for more details.

Related Actions

  • Search and locate based on the node name source.
  • Choose whether to display metrics and scores.
What are metrics and scores?
  • Metrics: duration time display
  • Scores: Model quality evaluation scores, descriptions, etc.

Preview

Includes node type, node name, execution duration, token consumption, execution status, and execution details.

Status Display

Execution Status (status) Possible Causes
Success: Success
Error: Error Agent/LLM call timeout, tool returned exception, etc.
Warning: Warning Output content triggered sensitive word detection but execution was not interrupted, etc.

Execution Details

Displays the Input and Output of the application. The Input contains all elements used to construct the final request sent to the model, while the Output shows the result generated from the input.

For data ingested via LangChain OpenTelemetry, the system parses gen_ai related attributes into Input, Output, and Token information. If the original data does not provide the corresponding attributes, the page displays empty fields.

Property Information

Field information related to the current Span or Trace.

Using the + button in the upper-right corner, you can bind built-in views to the current detail page for dedicated display of Agent/LLM data.

  1. Click the Add button.
  2. Select a related field.
  3. In the Bind Built-in View panel that appears, select the view to associate.
  4. After binding, the new tab appears in the tab bar.

Further Reading

You may also be interested in:

Feedback

Is this page helpful?