IFRS9 Credit Risk Workflow Failure – Jobflow Execution Error

[Musheer1]

We encountered a failure in the IFRS9 Credit Risk workflow. The jobflow did not complete successfully due to validation errors and missing data objects. Key observations:

• Validation failure on FRS9_STG.ECL_DETAIL.
• Missing source cube for scenario run (HTTP 404, errorCode=10014).
• Uninitialized macro variables (INSTTYPE, CURRENCY, PD_SEGMENT_ID, LGD).
• DataObjectVerificationException during jobflow execution.

Environment: SAS 9.4 TS1M7, IFRS9 Solution v10.2020
Cycle: Exim_IFRS9_ECL_Cycle_V2_10557_2025-09-30_EXIM
Instance Key: 264278757

Request:
Please advise on:

• Steps to resolve DataObjectVerificationException.
• Best practices for validating IFRS9 staging tables and cube setup.
• How to handle missing source cube errors in IRM scenario runs.
  
  

[Musheer1 · Accepted solution]

Hello everyone,

I wanted to share the solution provided by SAS Technical Support for an intermittent failure we were experiencing in our IFRS9 (Credit Risk Analysis) cycle runs.

The Original Problem: The SAS Model Implementation (MIP) job was successfully completing, but the MIP mid-tier was incorrectly marking the job as failed due to an execution timeout. The logs suggested this was likely due to a very slow response from the workspace server (sometimes pausing for 35+ seconds before executing code).

The Solution Provided by SAS Technical Support: The immediate fix was to adjust the mid-tier's timeout threshold to prevent it from prematurely marking the job as failed.

This requires modifying the configuration file: Lev1/Web/WebAppServer/SASServer11_1/sas_webapps/sas.modelimplementationmanagement.war/WEB-INF/spring-config/app-core-config.xml

We modified the following section, specifically increasing the timeout value (the first constructor argument) from the default of 60 seconds to a higher value (e.g., 120 seconds) to accommodate the delay:

<bean id="executionTaskCleanupService" class="com.sas.risk.modimp.services.system.ExecutionTaskTerminationServiceImpl">
        <constructor-arg type ="int" value="120" /> 
        <constructor-arg type ="int" value="2" />
</bean>

Result: After making this change and restarting the server, the MIP mid-tier failure is resolved, and the IFRS9 cycle now completes successfully without being marked as failed.

Current Status (Ongoing Performance Issue): While the job now completes successfully, we are currently troubleshooting a severe performance degradation. The cycle previously took 10–15 minutes, and now it takes 30+ minutes. The consensus is that the underlying issue of the slow workspace server still needs to be fully investigated and resolved.

We will share updates on the root cause analysis once that investigation is complete.

Hope this helps anyone dealing with similar MIP timeout errors!

Best regards,
Musheer Emad.