Dashboard

The Dashboard opens after sign-in and summarises HiveTrace traffic. Message and token metrics update as new data arrives without requiring a page reload.

Summary metrics and charts

Messages

• Total — requests and responses recorded over all time.
• In the last 24 hours — messages received during the latest rolling 24-hour period.

If traffic is expected but the daily value does not change, check the API token, application, and integration method.

Violations

• Total — messages for which at least one safety check triggered.
• In the last 24 hours — violating messages during the latest 24 hours.
• Violation summary chart — compares messages with and without violations.

The metric counts messages rather than triggered rules; one message can contain multiple violations. Hover over a bar for its category and exact count. The chart period is fixed to All time in version 4.3.

Token consumption statistics

• Total tokens — total recorded token consumption.
• Tokens today — consumption since the start of the current calendar day.
• Average tokens per request — average consumption per request.
• Total violations — number of token-threshold violations.
• Token usage analytics — percentage distribution of low, high, and critical violations.

The pie chart shows violation severities, not the ratio of prompt to completion tokens. Hover over a segment for its severity and share. Its period is fixed to All time in version 4.3.

Latest alerts

• ID — unique alert identifier.
• Alert type — trigger, such as token usage or a DataClean violation.
• Channels — configurations through which the notification was sent.
• Date and time — creation time.
• Direction — in for a user request or out for an LLM response.
• Severity — low, high, or critical.
• Row menu — actions for the selected alert.

Go to all alerts opens the complete alert log. Column settings hides or restores columns; it does not delete data or change alert configurations.

Review workflow

1. Compare the latest 24-hour metrics with normal traffic.
2. Assess the share of violating messages.
3. Review token-violation severity distribution.
4. Inspect recent alerts, starting with critical and high.
5. Open the full alert log or session analytics for investigation.

Values in the screenshots are example dev-environment data and are not recommended thresholds.