Problem

The Confluence Data Pipeline Export will fail with a NullPointerException if the CONTENT table contains records with CONTENTTYPE = 'USERINFO' and USERNAME IS NULL. The conditions that cause the USERINFO records to exist with a NULL USERNAME are unknown.

The following exception will be found in the atlassian-confluence.log file:

2022-09-12 17:00:38,215 ERROR [data-pipeline-export-executor-0] [insights.core.service.DefaultDataExportOrchestrator] lambda$null$5 Failed writing entities to file - processId: 25 java.lang.NullPointerException  
  
at com.atlassian.business.insights.confluence.extract.UserToLogRecordConverter.convert(UserToLogRecordConverter.java:37)    

at com.atlassian.business.insights.confluence.extract.UserLogRecordStreamer.toLogRecord(UserLogRecordStreamer.java:56) 

The Confluence database schema does allow NULL values in CONTENT.USERNAME.

Suggested Solution

This suggestion is to harden this aspect of the Confluence Data Pipeline Export and instead of failing the export: 

• Log a warning of the condition
• Skip the bad record and proceed with the export