In some customer environments after upgrading to 15100, High CPU utilization and Out of Memory errors due to insufficient Heap space were seen.
1. High CPU utilization error (SD-132540 released in 15110 build)
In 15100 build version, we have introduced Zia Parser feature which is enabled by default in IT helpdesk portal and disabled in other available portals. For this feature, data training will occur during service startup for all portals (even when Zia parser is disabled). This is to try out "Test Zia Parser" option inside the Zia parser feature. However, in some big environments, training in available portals triggered High CPU usage.
This has been fixed in 15110 build - Training during service startup will occur only for Zia parser feature enabled portals. This training will occur only once and the trained data will be stored in a file. Post this, during every service startup, training will not occur due to this stored data file.
2. OOM error while initializing the Zia parser (SD-134060 which is to be released in 15250 build)
For customers in 15100 build, use the attached fjar file and execute the below queries and restart the service once.
For customers in 15110 - 15130 build, execute the below queries and restart the service once.
update ai_module set action_enabled=false where actionname='parser';
delete from Schedule where SCHEDULE_ID in (select s.SCHEDULE_ID from Schedule s inner join AI_Module am on am.ACTIONNAME = 'parser' where s.SCHEDULE_NAME like CONCAT('AITrainingSchedule', CAST(am.MODULEID AS VARCHAR)));
Customers migrating from 15100 to build version 15130 or above, Zia parser will be disabled in all portals. If customer enabled Zia parser post 15130 migration, use the above query to disable it and fix the OOM error.
For customers in 15140 build or above, execute the below queries to disable zia parser and parser training schedules and restart the service once to fix the OOM error.

update zia_module_prediction_mapping set enabled=false where predictionid in (select id from zia_predictions where internalname like 'parser');

DELETE FROM Schedule WHERE SCHEDULE_ID IN ( SELECT s.SCHEDULE_ID FROM Schedule s INNER JOIN zia_module_prediction_mapping zmpm ON s.SCHEDULE_NAME LIKE CONCAT('AITrainingSchedule', CAST(zmpm.id AS VARCHAR)) INNER JOIN zia_predictions zp ON zmpm.predictionid = zp.id WHERE zp.internalname = 'parser' );