Steps to Reproduce
The UsageAnalytics Count Persistor scheduled job on an instance, which runs daily, can cause high memory utilization depending on the amount of data that is being collected.
If you are affected by this issue, the UsageAnalytics Count Persistor will get picked up by a worker at its daily run time and the memory will climb quickly as the job begins processing and runs the node out of memory.
Check stats.do for the job running and memory utilization. For example:
Max memory: 1980.0
In use: 1878.0
Free percentage: 5.0 --->> refresh the stats page and you will see the "In use" continue to climb as the job processes
Current job: UsageAnalytics Count Persistor
Job started: Fri Jun 17 03:08:27 PDT 2016
Job duration: 0:00:53.802
Total jobs: 17
Mean duration: 0:00:03.729
Set the Trigger Type to OnDemand to disable the UsageAnalytics Count Persistor scheduled job (/sys_trigger_list.do?sysparm_query=GOTOname%3DUsageAnalytics%20Count%20Persistor)
Restart the affected node.
Disable the problematic count config records for the Service Catalog data collection DEF1000110 and DEF1000112.
- Navigate to /usageanalytics_count_cfg_list.do?sysparm_query=definition_id%3DDEF1000110%5EORdefinition_id%3DDEF1000112
- Set each record to Active = false.
Re-enable the UsageAnalytics Count Persistor scheduled job:
- Set Trigger Type = "Daily"
- Set Run Time = 1am
Related Problem: PRB684790