VectorChat Docs

Monitoring Executions

Track execution history, review results, and troubleshoot scheduled agent runs.

VectorChat records every scheduled execution so you can track results, identify issues, and understand how your agents are performing over time.

Execution history

Each schedule maintains a history of all its executions. To view execution history:

  1. Navigate to Scheduled Agents in the sidebar.
  2. Click on a schedule to open its detail page.
  3. The execution history is listed in reverse chronological order.

Each execution entry shows:

  • Status — The current state of the execution.
  • Start time — When the execution began.
  • Completion time — When the execution finished (if applicable).

Execution statuses

StatusMeaning
RunningThe agent is currently processing the prompt.
CompletedThe execution finished successfully.
Pending ApprovalThe agent proposed actions that are waiting for review.
ApprovedThe execution was approved and all actions were executed.
RejectedThe execution was rejected and no actions were executed.
FailedThe execution encountered an error.

Viewing execution details

Click on any execution to see its full details:

Rendered prompt

The exact prompt that was sent to the agent, with all template variables replaced with actual values. This is useful for verifying that variables like {today} and {yesterday} resolved correctly.

Agent response

The full text response generated by the agent during the execution.

Pending actions

If the execution included tool calls, you can see each proposed or executed action with:

  • The tool name
  • A summary of the action
  • Detailed parameters
  • Execution status (pending, executed, or failed)
  • Any error messages

Troubleshooting

Executions are failing

Common causes:

  • Agent is disabled — The agent must be enabled for scheduled runs to work. Check the agent's status on its detail page.
  • Tool authentication expired — If the agent uses tools with OAuth, the token may have expired. Re-authenticate the integration.
  • Prompt issues — Review the rendered prompt to check for issues with template variables or unclear instructions.

Agent is not producing expected results

  • Review the rendered prompt — Check that template variables are resolving to the expected dates and times.
  • Test manually — Start a conversation with the same agent and send the same prompt to see how it responds interactively.
  • Check skill assignments — Verify that the agent has the correct skills attached for the task.
  • Review tool access — Confirm the agent has the necessary tools connected and authenticated.

Schedule is not firing

  • Check if paused — A paused schedule will not fire. Look for the "Paused" badge on the Schedules page.
  • Verify next run time — The schedule detail page shows the calculated next run time. Confirm it is in the future and reasonable.
  • Check timezone — Times are configured in Central European Time by default. Make sure the scheduled time accounts for the correct timezone.

Schedule status badges

On the Schedules list page, each schedule shows a status badge:

  • Active — The schedule is running normally.
  • Paused — The schedule is temporarily stopped.
  • Running — An execution is currently in progress.
  • Pending Approval — The latest execution needs your review.
  • Failed — The latest execution encountered an error.