Applies to Product – Dynamics 365 Finance
What’s happening?
Batch jobs are ending with errors, but no error details are logged. Additionally, sporadic crashes of BatchAOS with OutOfMemoryException are reported.
Reason:
- The batch job errors are occurring due to the simultaneous execution of multiple custom batch jobs, which leads to an out-of-memory error. This indicates that the system could not handle the execution of these jobs efficiently.
- During the execution period of the batch job, the AOS server and batch AOS server were restarted, which may have contributed to the lack of error logging.
- Review the custom batch jobs to ensure that the data sources do not contain null values.
- Monitor the system's memory usage during batch job execution to identify potential problems.
Additionally,
- Optimize Data Storage:
- Regularly archive old data to reduce the volume of live data.
- Use data archiving features to move historical data to long-term storage.
- Optimize Queries:
- Use Microsoft SQL indexes to speed up long-running data queries.
- Avoid complex queries and use out-of-the-box functions whenever possible.
- Minimize Customizations:
- Keep customizations to a minimum and ensure they are efficient.
- Avoid unnecessary custom code that can lead to memory leaks.
- Apply right LCS Subscription Estimator:
- Use tools in Lifecycle Services (LCS) to monitor performance and identify potential issues.
- Regularly update the LCS Subscription estimator to keep environment sizing updated - refer - https://learn.microsoft.com/dynamics365/fin-ops-core/dev-itpro/lifecycle-services/subscription-estimator#edit-the-estimate-for-multiple-implementation-projects.
- Manage Concurrent Users:
- Schedule resource-intensive tasks during off-peak hours to reduce the load on the system.
- Implement user load balancing to distribute the workload evenly.
- Clean Up Obsolete Data:
- Regularly run cleanup jobs to remove unnecessary data such as executed orders and transaction history.