Task History

Task History is the cross-flow execution dashboard. Every flow run — whether it succeeded, failed, was replayed, or is currently in-flight — shows up here with its status, duration, and a link back to the source flow.

Why it's useful

• Find recent failures across any flow without drilling into each one
• Audit activity during a specific time window
• Replay a failed run after fixing a step configuration
• Filter by status (pending, running, completed, failed) and by flow

Debugging a failed run

1. Open /admin/task-history and filter status to failed.
2. Click the flow name to jump to the flow detail.
3. Expand the failed execution row. Each step shows its input, its output, and — if it failed — the error message captured by the engine.
4. Fix the underlying issue (often a missing field in the input mapping or a stale connection). Click Replay on the row to re-run with the exact same trigger payload.

Replay vs test

Replay runs a past execution again, preserving the original trigger payload. Use it to verify that a fix works on the same inputs that caused the failure.

Test runs a flow with a payload you supply. Use it during development when you want to try arbitrary inputs without waiting for a real trigger event.

API

GET /connect/v1/task-history
    ?page=1&status=failed&flowId=<uuid>

GET /connect/v1/flows/:id/executions/:executionId/steps
POST /connect/v1/flows/:id/executions/:executionId/replay

Permissions

task_history.view gates the dashboard and the execution inspection endpoints. Replay additionally requires flows.write on the source flow.