Applies to Product - FnO Platform
What’s happening?
The customer reported issues related to the wait type "Parallelism" and abnormal terminations of batch jobs without producing expected outputs. Specifically, timeout errors occurred during processing tasks involving multiple threads and entries.
Reason:
The underlying cause of this appears to be related to the handling of parallelism in the processing tasks, which may lead to timeout errors when the system is under heavy load or when there are resource constraints.
Resolution:
- For the timeout error during processing tasks, it is recommended to monitor the wait statistics for the "PARALLELISM" wait category. This can be done using the following query: | where originalEventTimestamp between [start_time] .. [end_time] | where wait_category == "PARALLELISM" | project interval_id, wait_category, max_query_wait_time_ms | render columnchart | where max_query_wait_time_ms == [specific_value]
- For batch jobs that experience abnormal terminations, further investigation is needed to identify the specific error messages or conditions that led to the failure. If the error messages are not available, additional logging or monitoring may be required to capture the necessary details for troubleshooting.
- This might occur due to some crashes or memory issues caused by particular batch class . So we should verify the class ownership and investigate with concerned team.