Upgrading to Enterprise Search 7.10.0? See Upgrading & migrating.
- New programming language clients for Python and Ruby.
- We have a new and improved License Management page. Start a trial, downgrade, and see trial status, all from one clean and easy location. You can get there via the top right menu of Enterprise Search Home, where you choose between products.
- Official Enterprise Search Docker images are now based on CentOS 8.
- Embedded filebeat has been upgraded to the latest version.
- Embedded Jetty web server has been upgraded to the latest version.
- Improves resiliency of upgrades from Elastic App Search.
- Improves resiliency of preflight checks when Enterprise Search is starting up (longer retries, etc).
- Improved error logging to always include a detailed error message for each exception.
bin/enterprise-search --reset-authcommand for situations when console echo is disabled.
- Java GC logging support fixed for deployments using Java 9 or later.
Removed duplicate logging to
/var/log/messageson Linux systems using systemd.
- Significantly increased all Elasticsearch query timeouts to help with degraded results being returned from overloaded App Search deployments.
- API requests with PUT or DELETE method using HTTP/1.0 used to fail with an HTTP 400 error.
- Slack is now available as a private, remote content source!
- Salesforce Sandbox is now available as an organization content source.
- Microsoft SharePoint Online and OneDrive now support document-level permissions.
- You can now enable and disable access to specific private content sources via the Security menu.
- We have improved our analytics story. We capture more events, enough that you can populate a useful Kibana dashboard. Read our blog.
- Tidied up and simplified a set of navigation items and introduces Elastic Amsterdam styling.
- Confluence Server will now show large thumbnails.
- Clearing Group filters will no longer send you to the root domain.
- Dropbox will no longer 500 if it detects the incorrect access scope.
- The "Updated by" trigger will work now, even with lots of users.
- The 404 page had some minor errors. Ironic, and fixed!
- Private Sources can no longer be configured with a Basic license. You will need a Platinum license, as intended.
Subextractors were failing on Docker due to a missing dependency. We have added it:
- Google Drive has much improved resilience.
- Thumbnails will cause large deployments to OOM. We will now check for at least 4GB of memory and disable them when only 2GB or less is available. We will improve memory handling of Thumbnails so that they work with defautl 2GB deployments in the near future.
- Remote content source facet counts should again be accurate.
- We will now ignore transient fails after retry when indexing, so that the rest of the job can finish as expected.
- Google Drive no longer indexes your trash files.
- Many more general stability and quality fixes.