Synchrony is slow

[DerivcoITServer]

We have been experiencing slow performance when updating/saving pages with synchrony.  We often the following snippet in the logs:

2017-03-28 14:18:38,565 WARN [http-nio-8443-exec-56] [confluence.util.profiling.DurationThresholdWarningTimingHelperFactory] logMessage Execution time for publishing event com.atlassian.confluence.event.events.content.page.synchrony.ContentUpdatedEvent@3c17e0a5 took 18707 ms (warning threshold is 5000 ms)
-- referer: https://confluence/pages/resumedraft.action?draftId=34587494&draftShareId=73035f53-587a-4859-bc7b-09a3089e6f7d | url: /rest/api/content/34587486 | traceId: d734339b1217336c | userName: Altonx@abc.com
2017-03-28 14:18:38,566 WARN [http-nio-8443-exec-56] [confluence.util.profiling.DurationThresholdWarningTimingHelperFactory] logMessage Execution time for post-commit task com.atlassian.confluence.pages.DefaultPageManager$$Lambda$1189/1118956715@3bc26a6c took 18708 ms (warning threshold is 5000 ms)
-- referer: https://confluence/pages/resumedraft.action?draftId=34587494&draftShareId=73035f53-587a-4859-bc7b-09a3089e6f7d | url: /rest/api/content/34587486 | traceId: d734339b1217336c | userName: Altonx@abc.com
2017-03-28 14:18:38,579 WARN [http-nio-8443-exec-56] [atlassian.confluence.cache.TransactionalCacheFactoryCleaner] lambda$cleaner$2 Forcibly unbound thread-local transactional caches prior to request execution: [com.atlassian.confluence.core.CachingInheritedContentPermissionManager.getInheritedContentPermissionSets(), com.atlassian.confluence.user.crowd.CachedCrowdUserDao.USER_CACHE, com.atlassian.confluence.user.crowd.DefaultApplicationCache]
-- referer: https://confluence/pages/resumedraft.action?draftId=34587494&draftShareId=73035f53-587a-4859-bc7b-09a3089e6f7d | url: /rest/api/content/34587486 | traceId: d734339b1217336c | userName: Altonx@abc.com

How do we stop this and increase the performance?

[DerivcoITServer]

This has been resolved by adding the following index for postgres.

CREATE UNIQUE INDEX e_h_p_s_idx ON "EVENTS" (history, partition, sequence);

https://jira.atlassian.com/browse/CONFSERVER-45745

[DerivcoITServer]

We see times exceeding 45 seconds and we get these messages exceeding its threshold on every page update.