Symptoms

Confluence experiences slow performance at regular intervals. Thread dumps, taken across time, show this thread:

...
	at $Proxy29.optimizeIndex(Unknown Source)
	at com.atlassian.confluence.search.lucene.IndexOptimizer.executeJob(IndexOptimizer.java:20)
	at com.atlassian.confluence.setup.quartz.AbstractClusterAwareQuartzJobBean.surroundJobExecutionWithLogging(AbstractClusterAwareQuartzJobBean.java:64)
	at com.atlassian.confluence.setup.quartz.AbstractClusterAwareQuartzJobBean.executeInternal(AbstractClusterAwareQuartzJobBean.java:47)
...

Cause

The index optimization job is running and causing a backlog.

Diagnosis

Add debug level logging for the following class:

confluence.search.lucene.DefaultConfluenceIndexManager

Watch at 03:00:00 and 15:00:00 for these threads. If you're noticing long delays until the next DEBUG level instance of this class, it may be a culprit.

Alternatively, watch for the thread above in your thread dumps.

Resolution

By default, the indexOptimizerTrigger is set to run at 3am and 3pm daily. Configure the index optimization to run once a week, or less often.