Known issues with high CPU utilization
Check the Insight system requirements before you try any of the following solutions.
In rare cases the CPU utilization might reach over 90-95% when Insight is enabled. This can lead to Jira crashing. There can be different reasons the CPU utilization is unusually high:
- Insight doesn't have enough dedicated power. Check the Insight system requirements and note that these numbers are measured only for Insight to run smoothly. More add-ons with scheduling tasks raise these numbers.
- Insight gadgets in dashboards. Many users having multiple insight gadgets in place can cause high CPU utilization as the gadgets send REST requests (each individually) to Insight. This causes the JVM to be fully utilized. We recommend setting up a shared filter or shared reports instead.
- Too many automation rules running at the same time. Setting up a long idle interval can help with this.
- The GC overhead is exceeded. This can be fixed through setting up garbage collection settings in Jira.
- In Data Center environments the database gets filled up very quickly. This causes the database to go out of memory and is related to how the Jira API handles messages in the clustermessage table in order to replicate the Insight data across all nodes. Configure global Jira settings to solve this issue.
- The database connections aren't enough. Increasing Insight parallelism might require increasing the database pool.
- Importing too many objects in one go. Try to filter the objects you're importing with IQL (Filter Data source using IQL) or only syncing one object type at a time. This depends on how many objects you've loaded in the system and how many objects you're importing from the data source.
- Exporting an object schema with a big number of objects and/or a lot of attachments can cause the server to go out of memory. We recommend using Object schema import type instead with the recommendation from number 7.
- Having too many indexed object attributes in the memory. You can turn turn this option off for certain attributes which will help to lower the memory footprint consumption.