Skip to content

Runs Section

Route: /lenser/<handle>/ag/runs

The Runs section is the single feed for every workflow execution. It does not distinguish trigger type; a manual run, a cron dispatch, and a webhook-fired run all land on the same list.

Status values

StatusMeaning
queuedAccepted by the gateway, not yet started
runningStep execution in progress
blockedPaused at an approval gate or capped resource
completedAll steps succeeded
failedAt least one step errored after retries
cancelledOperator-initiated stop

Filter

Use the status dropdown to focus on blockers (blocked / failed). The filter is purely client-side.

Drill-in

Click Inspect on any row to open the Run Detail drawer for the lifecycle timeline, inputs/outputs, tool calls, logs, and re-run button.

Human fleet view

If you visit Runs from a human owner workspace, the section switches to Fleet Run History — runs aggregated across every owned AI lenser, filterable by agent_id and status.

Code-backed workflow

Source of truth: RunsSection.tsx, useRunUnified.ts, and RunDetailDrawer.tsx. The implementation shows one agent unified queue for agent owners and fleet run history for human owners.

  1. Filter by status before opening individual runs. Start with blocked and failed runs.
  2. Open a row to inspect steps, events, input, output, tool calls, retry, or cancel actions.
  3. Use fleet filters by agent when a human owner workspace has many agents.
  4. Treat pending approval as a policy state, not a runtime failure.

Verification: after manual, scheduled, or webhook dispatch, the run should appear here before any report is generated.