Helpful Information Regarding OE and CDC Logging

TABLE OF CONTENTS

• Purpose of OE Logging
• Impact of Enabling OE Logging
• Is OE Logging the same as CDC Logging?
• Why might a task fire but fail to evaluate a CDC-tracked change?
• Can OE be triggered too closely together and miss updates?

This article provides some useful information regarding OE and CDC logging.  Clients may be asked by the Client Success team to enable these features in order for the Client Success team to capture data needed for triage.

Purpose of OE Logging

OE logging is mainly used to troubleshoot issues with OE tasks.  It records detailed information about task execution, including triggers, expressions, and condition failures.  In addition, it helps analyze why automated tasks may not have fired, providing visibility into the overall OE workflow execution.

Impact of Enabling OE Logging

Turning on OE logging significantly increases log volume. It generates highly detailed logs for every evaluated condition, making them more verbose than standard Empower logs. This can introduce some performance overhead. For this reason, OE logging is recommended only for short-term troubleshooting and should be disabled once the issue is diagnosed.

Is OE Logging the same as CDC Logging?

No, OE logging and CDC logging are not same.  The differences are as such:

• CDC Logging captures all data-level changes at a granular level and retains them for a period of time. This has historically raised performance concerns.
• OE Logging captures the logic flow within the Orchestration Engine, task triggers, condition evaluations, and reasons why a task did or did not execute.

Unlike CDC logging, OE logging does not record the full change history, so it does not carry the same performance impact.

Why might a task fire but fail to evaluate a CDC-tracked change?

Our review suggests several possible scenarios:

1. Trigger vs. evaluation timing– The task may fire correctly, but during evaluation:
   • A required field may not yet be committed
   • The value may not meet expression criteria
   • The system may retrieve a cached (previous) value
2. Replication delays – CDC fields update at the database level, but OE reads through the ODS/CDC pipeline. If replication is delayed or skipped, OE may fire but evaluate stale data.
3. Rapid task execution – Tasks triggered in quick succession can overlap. OE queues tasks sequentially, but closely timed triggers may cause evaluations before dependent updates are fully committed.

Can OE be triggered too closely together and miss updates?

OE does not lose updates but closely timed triggers can create race-condition-like timing issues: 

• The trigger detects a change immediately
• The evaluation step occurs before related fields are fully updated or replicated 

This is rare but possible under high activity conditions or when multiple systems write to the loan simultaneously. We are considering this as part of our investigation.