Page Restriction design does not scale well due to re-indexing of all sub-pages

Page restrictions, when applied to large spaces, can cause significant performance problems (instance becomes inoperable) due to the need to reindex every page (including every attachment). We've only recently begun to diagnose this in support, but we suspect it's been a problem for some time (based on the high frequency of incidents since learning what to look for).

http://confluence.atlassian.com/display/DOC/Page+Restrictions+Performance+Considerations

Implementation notice
The solution to this issue is to batch the contentns of the index queue. That way Confluence should remain responsive while re-indexing. The batch size defaults to 1500 so it should be large enough that most instances do not need to adjust it. However should you wish to adjust the batch size it is possible by passing -Dindex.queue.batch.size=1500 (replacing 1500 with your desired batch size).