Views:

Applies to Product - FnO Platform

What’s happening?
Customers are experiencing blocking issues in the production environment, particularly related to batch jobs that are stuck in executing status. The overall performance of the system is poor, and specific queries are causing delays.


Reason:
The blocking is primarily due to long-running SQL transactions and customized batch jobs that are blocking other processes. In particular, processes have been found to be suspended without completion, leading to a blocking tree that affects overall system performance.


Resolution:

  • Investigate the current blocking tree using SQL actions in the Lifecycle Services (LCS) to identify suspended processes.
  • Kill the blocking process (e.g., process 645) in LCS to allow other batch jobs to resume.
  • Review the telemetry data for the blocked processes to understand the duration of transactions and identify any long-running queries.
  • Engage with the customer to review their customization logic, as it may be causing long transaction durations.
  • If necessary, suggest splitting large batch jobs into smaller tasks and utilizing multi-threading to improve performance.
  • Confirm the current status of batch jobs with the customer and determine if further actions are needed, such as downgrading the case severity or upgrading SQL server resources temporarily to enhance performance.
If this persists or if there are no clear steps to resolve the blocking, further investigation into the customization logic may be required.