Dashboard

Guardrails and failures

Permissions, unattended approvals, limits and recovery after automatic failures.

Permissions and approvals

Action or boundaryRule
Inspect workflows, nodes and tracesDashboard viewer, editor or owner.
Build, run, change triggers or reveal secretsDashboard editor or owner.
Assistant proposes a workflowOne approval covers the whole program and branches; credentials, activation and real tests have separate approvals.
Agent stepFull auto: tools normally set to Ask execute without asking. Off tools remain blocked.
Tool step identityAccount-owned connection or credential profile, selected per step; never a member's personal account.
WorkspaceWorkflows reach the common tree; protection blocks and system-entry rules still apply.
App functionsA manager's workflow save authorizes the apps used; callers share that access. See app access.

Rule — Workflows cannot pause for tool approval. Grant workflow agents only tools you trust to execute unattended.

Limits

LimitValue or behaviour
Concurrent runs of one workflow1; manual or scheduled starts during a run are refused. External deliveries queue.
Steps20 across the complete workflow.
Declared inputs20.
Nested conditions8 levels.
Rules in one condition10.
Prompt output fields20.
Repeat capRequired, 1–20 items; exceeding it fails the run.
Repeat executionReads and prompts parallel; writes and agents sequential. Item failures are collected by default.
Minute interval5–1,440 minutes.
Hour interval1–24 hours.
Recent deliveriesLatest 50; delivery data retained for 30 days.
Sample data16,000 bytes.
Webhook request body256,000 bytes.
Webhook rate120 requests per minute per webhook.
Queued external deliveries200 per workflow.

Automatic failures

SituationResult
Failed scheduled, webhook or event runFailure email to the workspace owner.
Third consecutive automatic failureWorkflow switches off; a final email reports the disablement. The UI shows Off after failures.
Failed manual run or user stopDoes not count toward automatic failures.
Successful run, including manualEnds the failure streak.
Save or switch the workflow onRestarts the failure count.
Account runs out of SparksScheduled runs are skipped and incoming events dropped, without workflow failure emails.

For recovery, inspect the failed step in Runs, correct the workflow and use Switch back on, or enable it in Settings. Saving resets the count but does not itself reactivate a disabled workflow.

Costs and retry safety

OperationWhat to account for
Prompt or agent workUses Sparks; each repeated agent item is a complete agent run.
Sample testReal actions and normal Spark usage.
List's daily projectionEnabled schedules on active workflows only; external and manual traffic is excluded.
App write with an uncertain resultThe change may have happened; inspect the app before retrying.
New app publication changes a function contractSave the workflow again and review arguments and dependent outputs before retrying.