Debug with logs and traces - Veryfront

Use Logs to inspect runtime failures. Use Traces to inspect agent runs. Use Metrics to inspect measurements over time.

Runtime issues

Open Logs.
Find the log entry or time window that matches the issue.
Check the active deployment, server, or route involved.
Use Code to inspect the source linked to the failure.

Load more reads earlier log entries in the same time window.

Agent issues

Open Traces.
Select the agent run that produced the unexpected result.
Review messages, tool calls, and timing.
Use Tools or Agents to adjust the capability that failed.

Metric issues

Open the Metrics panel.
Select the metric and time range that match the issue.
If the panel shows a message, follow the matching row.

Message	What to do
Too many metric series	Add a label filter or choose an aggregation.
The metric result is too large	Add a label filter, choose an aggregation, enter a larger Step, or choose a shorter Time range.
Could not load metrics	Reload the browser page and open the Metrics panel. Then enter a larger Step or choose a shorter Time range.
Too many metric requests	Wait a moment, then select Reload.

The panel never draws a partial chart. Could not load metrics often means the time range and step need more than 11,000 points in one series.

Outside Studio

Agents and MCP clients read the same data with the list_project_logs, query_project_metrics, and list_project_alerts tools. The Observability API describes the limits and cursors.