Monitor Smart Flow executions
After you publish a Smart Flow, monitoring confirms it runs reliably. SWS Agent gives you four surfaces: Run history on the recipe, Execution Viewer, Trigger Events debugging, and System Ops for platform health.
Video
Who can do this
Admin and automation owners. Members typically see only outcomes on contact timelines, not execution internals.
Before you begin
- At least one saved workflow (Draft or Live)
- A contact ID for trigger debugging (copy from contact URL:
/contact/:contactId) - For failed runs: note approximate time and contact affected
What you'll accomplish
- Confirm successful runs and identify failed steps
- Diagnose why a trigger did not match a flow
- Review DLQ/retry trends before they affect many contacts
Run history and Trace (recipe panel)
Path: /smart-flows/recipe/:workflowId → right panel
Run history tab
- Heading: Recent runs
- Each row: timestamp, status, contact reference
- Failed rows show Error: message or Run failed — open Trace for step details.
- Empty: No automation runs yet — Once this automation starts running, you'll see its history here.
Trace tab
- Select a run from the Run dropdown.
- Review Execution outcome: SUCCESS, PARTIAL, SKIPPED, or FAILED.
- Expand Step status sections:
- Trigger & completion
- Filters (if conditions blocked enrollment)
- Domain actions (each action step)
Skip reasons you may see:
- Filters blocked this action
- Action skipped
Verify: A successful production run shows SUCCESS with green step statuses.
Execution Viewer (snapshot inspection)
Path: /smart-flows/executions/:workflowId
Open from dashboard Stats icon, recipe More actions → Execution history, or Advanced Builder Executions button.
Left panel — execution list
- Sort by Started At, Status, or Duration (Desc / Asc)
- Each row: status badge, execution ID, timestamp,
mode: live - Dead Letter Queue badge on failed retries
- Pagination: Prev / Next
Main canvas
- Execution Snapshot - Read Only
- Node graph shows what ran at each step (read-only—edit the flow in Recipe/Builder to fix issues)
Re-drive from DLQ
If banner shows Dead Letter Queue:
- Review failure reason on the snapshot.
- Fix the flow or integration.
- Select Re-drive Execution (shows Re-driving... while processing).
Save the workflow before opening Executions from Advanced Builder—unsaved changes prompt: Please save the workflow first before viewing executions.
Trigger Events (why did this not fire?)
Path: /smart-flows/trigger-events
Subtitle: Inspect workflow trigger lifecycle for a contact (last 100 traces, 24h retention)
Contact traces
- Enter Contact ID (placeholder: e.g. 550e8400-e29b-41d4-a716-446655440000).
- Select View traces (shows Loading…).
- Review lifecycle stages in order:
| Stage label | Meaning |
|---|---|
| Trigger received | Event entered the system |
| Contact events received | CRM event processed |
| Enqueued to queue | Queued for worker |
| Evaluating triggers | Matching flows |
| Workflow matched | A flow matched |
| Workflow filters did not match | Flow matched trigger but conditions failed |
| Workflow started / Workflow completed | Run lifecycle |
| Workflow failed | Run error |
| No workflow match | No flow had this trigger |
Empty states:
- Enter a contact ID and click View traces.
- No traces found for this contact. Traces are kept for 24 hours.
Workflow match diagnosis
When a flow should have run but did not:
- Enter Workflow ID (UUID from recipe Info tab → Automation ID).
- Confirm Workspace selection.
- Optionally reuse Contact ID above.
- Select Diagnose.
- Read Filter details (field → expected vs actual).
- Result: ✓ Would match or ✗ Would not match.
Verify: After fixing conditions, diagnosis shows ✓ Would match for the same contact.
System Ops (workspace health)
Path: /smart-flows/system
Open from dashboard card System Ops.
Page title: System Operations
Subtitle: Read-only provider health and workflow execution insight
Select Refresh to update metrics.
Metric cards
| Metric | Use |
|---|---|
| Total Executions | Volume |
| Success Rate | Overall health |
| DLQ Rate | Failures sent to dead letter queue |
| Retry Rate | Automatic retries |
| Avg Duration | Performance |
| DLQ Count | Items needing attention |
Sections
- Provider Health — Circuit Open vs Healthy; failure rate, cooldown
- Retry Trends (Last 14 Days)
- Per-Action Success Rate — table: Action Type, Total, Success, Success Rate
Use System Ops weekly; spike in DLQ Rate after an integration change often means expired credentials.
Verify your work
| Check | Expected result |
|---|---|
| Test contact run | SUCCESS in Trace |
| Production contact | Run appears in Run history within minutes of trigger |
| Trigger debug | Lifecycle reaches Workflow started |
| After fix | Per-Action Success Rate improves on next Refresh |
Tips
Copy Automation ID from recipe Info tab before opening support tickets—it speeds diagnosis.
Traces expire after 24 hours—debug trigger issues the same day they are reported.
If Filters blocked this action appears often, simplify conditions or document expected field values for reps.
Troubleshooting
| Symptom | Likely cause | Action |
|---|---|---|
| No runs in history | Flow still Draft or trigger never occurred | Toggle Live; fire test trigger |
| No workflow match | Wrong trigger type or inactive flow | Compare trigger catalog vs flow config |
| Workflow filters did not match | Condition field empty on contact | Fix data or relax filter |
| High DLQ Rate | External API failures (email, WA) | System Ops → Provider Health; reconnect integration |
| Legacy execution schema warning | Old run format | Re-run on current flow version; partial snapshot is normal |