Execution Trace¶
The Execution Trace provides a detailed, step-by-step view of a single orchestration execution. It shows exactly what happened at each step: inputs, outputs, timing, and errors.
Accessing a Trace¶
There are several ways to open an Execution Trace:
- From the Execution Log, click on any execution row
- From the Orchestration Runner, after running an orchestration, the trace appears automatically
- From an alert notification, click the execution link
Trace Layout¶
The Execution Trace displays a vertical timeline of steps, each shown as a card:
┌─────────────────────────────────────────┐
│ decode_vin ✓ 245ms │
│ API Call → NHTSA VPIC │
├─────────────────────────────────────────┤
│ extract_fields ✓ 12ms │
│ Transform │
├─────────────────────────────────────────┤
│ write_to_dms ✗ 1203ms │
│ API Call → DMS │
│ Error: 503 Service Unavailable │
└─────────────────────────────────────────┘
Each step card shows:
| Element | Description |
|---|---|
| Step name | The name you gave the step |
| Status icon | Checkmark (success), cross (failed), or skip icon |
| Duration | How long the step took in milliseconds |
| Step type | API Call, Transform, Conditional, or Error Handler |
| Target | For API calls, the connected system and operation |
| Error | For failed steps, the error message |
Step Detail¶
Click on any step card to expand it and see the full details.
Input Payload¶
The exact data that was passed to this step after all parameter mappings were resolved:
This is invaluable for debugging -- you can see exactly what values the parameter mapping produced.
Output Payload¶
The data returned by the step, available to subsequent steps via parameter mapping:
{
"status_code": 200,
"response": {
"Count": 1,
"Results": [
{
"Make": "Honda",
"Model": "Accord",
"ModelYear": "2003",
"BodyClass": "Sedan/Saloon"
}
]
}
}
HTTP Details (API Call Steps)¶
For API Call steps, additional HTTP-level details are shown:
| Field | Description |
|---|---|
| Method | HTTP method (GET, POST, PUT, DELETE) |
| URL | The full URL that was called |
| Request Headers | Headers sent with the request (credentials are redacted) |
| Response Status | HTTP status code |
| Response Headers | Headers returned by the external system |
| Response Body | The full response body (truncated if very large) |
Error Details (Failed Steps)¶
For failed steps, the detail panel shows:
| Field | Description |
|---|---|
| Error Type | The category of error (Network, Auth, Timeout, Application) |
| Error Message | A human-readable description of what went wrong |
| HTTP Status | The HTTP status code (if applicable) |
| Response Body | The error response from the external system (if available) |
| Retry Attempts | How many retries were attempted before the step was marked as failed |
Timeline View¶
The trace includes a visual timeline bar at the top that shows the relative duration of each step:
This helps you quickly identify:
- Which steps take the longest (performance bottlenecks)
- Whether there are gaps between steps (indicating processing overhead)
- The total execution time vs the sum of step durations
Comparing Executions¶
To compare two executions of the same orchestration:
- Open the first execution's trace
- Click Compare in the toolbar
- Select a second execution to compare against
- The two traces are shown side-by-side
Differences are highlighted:
- Steps that succeeded in one execution but failed in the other
- Duration differences (faster/slower)
- Input parameter differences
- Output data differences
This is useful for understanding why one execution succeeded and another failed.
Data Sensitivity¶
DIBOP respects PII classification when displaying trace data:
- Fields classified as PII in the Canonical Data Model are partially masked (e.g.,
j***@example.com) - Fields classified as Sensitive PII are fully redacted (e.g.,
[REDACTED]) - Non-PII fields are shown in full
Full Access for Debugging
If you need to see unmasked PII data for debugging, your platform administrator can grant temporary elevated access. This access is logged for audit purposes.
Copying and Sharing¶
Copy Execution Link¶
Click the Copy Link button to copy a direct URL to this execution trace. Share it with colleagues who have access to DIBOP.
Export Trace¶
Click Export to download the full trace as a JSON file. This includes all step inputs, outputs, and metadata.
Copy Step Data¶
In any step's detail panel, click the copy icon next to the input or output payload to copy it to your clipboard. This is useful for reproducing issues in external tools.
Troubleshooting with Traces¶
"Why did this step fail?"¶
- Click the failed step to expand it
- Check the Error Details section for the error type and message
- If it is an API Call, check the HTTP status and response body
- Common causes:
- 401/403: Credential issue → Manage Credentials
- 404: Wrong endpoint or resource not found → Check the connector configuration
- 500: External system error → Check if the external system is operational
- Timeout: The system took too long to respond → Increase the timeout
"Why did the mapping produce unexpected data?"¶
- Click the step where the data looks wrong
- Compare the Input Payload with what you expected
- Work backwards through the trace to find where the data diverged
- Check the parameter mapping expressions for typos or incorrect field paths
"Why is this execution so slow?"¶
- Check the Timeline View to identify the slowest steps
- Click on slow steps to see their duration breakdown
- For API Call steps, check the external system's response time
- Consider adding timeout limits or parallel execution for independent steps
Next Steps¶
- Execution Log -- browse all executions
- API Call Log -- drill into individual API calls
- Setting Up Alerts -- get notified of failures
- Parameter Mapping -- debug data flow issues