Issue Summary

The Lucene statistics for the type of directory RAMDirectory are written to the Jira log files as INFO level when this type of directory is flushed, which generates additional logging for an event that is not expected to be logged at INFO level.

It was found in the logs files analyzed (please note that this can vary for each deployment) that 25% of the Lucene statistics generated belonged to the flushing of RAMDirectory.

It is important to note that for any other type of index this log provides valuable statistics which could be use to proactively identify possible index performance degradation.

Steps to Reproduce

There are not steps to reproduce the issue.

Expected Results

Flushing an IndexWriter of a RAMDirectory should not be logged as INFO log level, however it could be considered as TRACE log level instead.

Actual Results

The below log lines appear very often in the Jira log files.

2019-11-12 15:40:17,805 catalina-exec-349 INFO anonymous 940x4070627x10 - [IP Address Sanitized],[IP Address Sanitized] /rest/api/latest/search [c.a.jira.index.MonitoringIndexWriter] [lucene-stats] flush stats: snapshotCount=20049, totalCount=2483576, periodSec=300, flushIntervalMillis=14, indexDirectory=null, indexWriterId=com.atlassian.jira.index.MonitoringIndexWriter@45ea1f81, indexDirectoryId=RAMDirectory@52bedaf6 lockFactory=org.apache.lucene.store.SingleInstanceLockFactory@79461c10

2019-11-12 15:40:17,805 catalina-exec-349 WARN anonymous 940x4070627x10 - [IP Address Sanitized],[IP Address Sanitized] /rest/api/latest/search [c.a.jira.index.MonitoringIndexWriter] Detected frequent flushes (every 14 millis) of lucene index which is below warning limit: jira.index.warn.flush.min.interval.millis=1000 millis. This may affect indexing performance. Please visit https://confluence.atlassian.com/x/w0VwOQ for more information.

Workaround

Adding and Setting the com.atlassian.jira.index.MonitoringIndexWriter package log level to ERROR.