Summary

The JOURNALENTRY table keeps growing or it is quite big (~1 million entries), but the Clean Journal Entries is not removing those items and there are only few thousands removed on each execution. 

Diagnosis

You can use the following SQL statement to check how many records are removed on each execution: 

SELECT * FROM scheduler_run_details where job_id LIKE '%JournalCleaner%' order by start_time desc;

And compare it with the actual number of items in the JOURNALENTRY: 

SELECT COUNT(*) FROM JOURNALENTRY;

There is a job called Clean Journal Entries which periodically clears journal entries that have already been processed to ensure that its size does not grow indefinitely

Cause

The JOURNALENTRY table tracks all content changes that needs to be indexed and every Confluence node keeps a journal file under <confluence_home>/journal  that store a number that corresponds to the last item that was indexed on the node from the JOURNALENTRY table. There is a job called Clean Journal Entries which periodically clears journal entries that have already been processed to ensure that its size does not grow indefinitely. 

By default entries older than two days are removed, but the life time of entries can be overriden with com.atlassian.confluence.journal.timeToLiveInMillis  system property.

Solution

TRUNCATE the content of the JOURNALENTRY table and rebuild your index from scratch as outlined in KB How to manually rebuild content index from scratch on Confluence Data Center without any downtime

Monitor the JOURNALENTRY table and if keeps growing along the days, you can adjust the com.atlassian.confluence.journal.timeToLiveInMillis system property to a more restrictive value: 

1. For example, if you want to reduce this life time to one single day, set this System Properties in your setenv.sh file as follows:

   CATALINA_OPTS="-Dcom.atlassian.confluence.journal.timeToLiveInMillis=86400000 ${CATALINA_OPTS}"

2. Restart Confluence.