Troubleshoot your live journey execution troubleshooting-execution
In this section, learn how to troubleshoot journey events, check if profiles entered your journey, how they navigate through it, and if messages are sent.
You can also troubleshoot errors before testing or publishing a journey. Learn how on this page.
If you are using inbound actions, learn how to troubleshoot them on this page.
Check that events are properly sent checking-that-events-are-properly-sent
The starting point of a journey is always an event. You can perform tests using tools such as Postman.
You can check if the API call you send through these tools is sent correctly or not. If you get an error back, it means that your call has an issue. Check the payload again, the header (and especially the organization ID) and the destination URL. You can ask your administrator what is the right URL to hit.
Events are not pushed directly from the source to journeys. Indeed, journeys rely on Adobe Experience Platform’s streaming ingestion APIs. As a result, in case of event related issues, you can refer to Adobe Experience Platform documentation for Streaming ingestion APIs troubleshooting.
If your journey fails to enable test mode with error ERR_MODEL_RULES_16, ensure the event used includes an identity namespace when using a channel action.
The identity namespace is used to uniquely identify the test profiles. For example, if email is used to identify the test profiles, the identity namespace Email should be selected. If the unique identifier is the phone number, then the identity namespace Phone should be selected.
Check if people enter the journey checking-if-people-enter-the-journey
Journey reporting measures people’s entrances in a journey in real-time.
If you are successfully sending the event but see no entrance in the journey, it means that something goes wrong between the event sending and the event reception in the journey.
You can start troubleshooting with the questions below:
-
Are you sure the journey where you expect the incoming event is in test mode or live?
-
Did you save your event before copying the payload from the payload preview?
-
Does your event payload contain an event id?
-
Did you hit the right URL?
-
Did you follow the Streaming Ingestion APIs payload structure, using the payload structure preview in the event configuration pane? See this page.
-
Did you use the right key-value pairs in the header of your event?
code language-none X-gw-ims-org-id - your organization's ID Content-type - application/json
Check how people navigate through the journey checking-how-people-navigate-through-the-journey
Journey reporting measures the progress of individuals inside a journey. It’s easy to identify where and why a person got stopped.
Here are a few things to check:
- Is it due to a condition excluding the person? For example, the condition is “gender = male” and the person is a woman. This check can be performed by a business user if the condition is not too complex.
- Is it due to a call to a data source not responding? When the journey is in test, this information can be seen in test mode logs. When the journey is live, an administrator can test direct calls to the data source and check the answer received. An administrator can also duplicate the journey and test it.
Check that messages are sent successfully checking-that-messages-are-sent-successfully
If individuals flow the right way in the journey but do not receive messages they should receive, you can check if:
- Journey Optimizer has correctly taken into account the request to send the message. Business users can access the message supposed to be sent and check if the time of the latest execution corresponds to the execution time of your journey. They can also check the latest API calls/events received.
- Journey Optimizer has successfully sent the message. Check the journey reporting to make sure that there are no errors.
In case of a message sent via a custom action, the only thing that can be checked during journey test is the fact that the call of the custom action’s system leads to an error or not. If the call to the external system associated with the custom action does not lead to an error but does not lead to a message sending, some investigations should be done on the external system’s side.
Understanding duplicate entries in Journey Step Events duplicate-step-events
Why do I see multiple entries with the same journey instance, profile, node, and request IDs?
When querying Journey Step Events data, you may occasionally observe what appears to be duplicate log entries for the same journey execution. These entries share identical values for:
profileID- The profile identityinstanceID- The journey instance identifiernodeID- The specific journey noderequestID- The request identifier
However, these entries have different _id values, which is the key indicator that distinguishes this scenario from actual data duplication.
What causes this behavior?
This occurs due to backend auto-scaling operations (also called “rebalancing”) in Adobe Journey Optimizer’s microservices architecture. During periods of high load or system optimization:
- A journey step event begins processing and is logged to the Journey Step Events dataset
- An auto-scaling operation redistributes workload across service instances
- The same event may be reprocessed by another service instance, creating a second log entry with a different
_id
This is an expected system behavior and is working as designed.
Is there any impact on journey execution or message delivery?
No. The impact is limited to logging only. Adobe Journey Optimizer has built-in deduplication mechanisms at the message execution layer that ensure:
- Only one message (email, SMS, push notification, etc.) is sent to each profile
- Actions are executed only once
- Journey execution proceeds correctly
You can verify this by querying the ajo_message_feedback_event_dataset or checking action execution logs - you’ll see that only one message was actually sent, despite the duplicate journey step event entries.
How can I identify these cases in my queries?
When analyzing Journey Step Events data:
-
Check the
_idfield: True system-level duplicates would have the same_id. Different_idvalues indicate separate log entries from the rebalancing scenario described above. -
Verify message delivery: Cross-reference with message feedback data to confirm only one message was sent:
code language-sql SELECT timestamp, _experience.customerJourneyManagement.messageExecution.messageExecutionID, _experience.customerJourneyManagement.messageDeliveryfeedback.feedbackStatus FROM ajo_message_feedback_event_dataset WHERE _experience.customerJourneyManagement.messageExecution.journeyVersionID = '<journeyVersionID>' AND TO_JSON(identityMap) like '%<profileID>%' ORDER BY timestamp DESC; -
Group by unique identifiers: When counting executions, use
_idto get accurate counts:code language-sql SELECT COUNT(DISTINCT _id) as unique_executions FROM journey_step_events WHERE _experience.journeyOrchestration.stepEvents.journeyVersionID = '<journeyVersionID>' AND _experience.journeyOrchestration.stepEvents.profileID = '<profileID>'
What should I do if I observe this?
This is normal system behavior and no action is required. The duplicate logging does not indicate a problem with your journey configuration or message delivery.
If you’re building reports or analytics based on Journey Step Events:
- Use
_idas the primary key for counting unique events - Cross-reference with message feedback datasets when analyzing message delivery
- Be aware that timing analysis may show entries clustered within a few seconds of each other
For more information about querying Journey Step Events, see Examples of queries.