A newer version is available. For the latest information, see the
current release documentation.
7.15.0 release notes
edit7.15.0 release notes
editUpgrading to Enterprise Search 7.15.0? See Upgrading & migrating.
Enterprise Search
editNew features
edit-
Adds the configuration setting
kibana.host
to specify the internal Kibana host that Enterprise Search communicates with.kibana.external_url
, an existing setting, now defaults to the value ofkibana.host
, but you can change this external URL for your deployments. Enterprise Search uses the external URL to render links to Kibana.
Bug fixes
edit- Clears reset password token after updating password so the same token cannot be used again. Prior to this fix, the token would eventually expire after 2 hours.
App Search
editNew features
edit- Promotes web crawler from beta to general availability (GA).
- Adds web crawler UI to Enterprise Search in Kibana.
- Adds process crawls, which re-apply updated crawl rules without requesting a full crawl.
- Adds web crawler UI for duplicate document handling.
- Adds web crawler API operations for URL validation and debugging.
- Adds web crawler operations to programming lanaguage clients.
- Improves curations. Promoted documents are about 15% faster, appear in and respect sorted results, and are returned in grouped queries.
- Adds a thumbnail image option to the Search UI page within App Search.
Breaking changes
edit-
Improvements to curations resulted in the following changes to the current behavior:
-
Positioned result scores will always be greater than organic results (previously they were always
1
). - Positioned results are returned in sorted queries (sorted with the rest of the results).
- Positioned results are returned in grouped queries.
-
Positioned result scores will always be greater than organic results (previously they were always
- The web crawler’s default user agent has been changed to "Elastic-Crawler" and will now include the full product version.
-
Fixes a known issue of Confluence Cloud’s, where attachment downloading via API was broken for OAuth apps.
To download attachments from Confluence Cloud, a new OAuth API Permission (scope) must be added to you Confluence Cloud OAuth Application:
Download attachment contents
.
Enterprise Search versions 7.15.0 and later are required to download attachments from Confluence Cloud. This is due to API changes introduced by Atlassian, the developer of Confluence Cloud.
- Updates the Confluence Cloud connector to work with rotating refresh tokens, which are mandatory with Atlassian cloud OAuth 2.0 apps after 2021-11-01.
After 2021-11-01, Enterprise Search versions 7.15.0 and later are required to connect to Confluence Cloud. This is due to backwards incompatible API changes introduced by Atlassian, the developer of Confluence Cloud.
Bug fixes
edit- Fix issue where the crawler domain validation would fail URL content check in case of redirect.
- Fix crawling pages with a relative base URL.
Known issues
edit-
App Search users with
dev
andeditor
roles do not have access to the web crawler. This will be addressed in an upcoming release. - Deleting an engine that has an active web crawl running can result in blocking future crawls from starting. This will be addressed in an upcoming release.
Workplace Search
editNew features
edit- Allows customizing the logo within the search experience.
-
Enables automatic query refinement for all content sources through the
automatic_query_refinement
API parameter. -
Adds an API operation to show the details for a content source’s automatic query refinement:
GET /api/ws/v1/sources/<source id>/automatic_query_refinement
. - Allows enabling and disabling of synchronization and synchronization features per content source (requires Platinum subscription).
- Adds a job management API operation that allows scheduling and interrupting content source synchronization jobs (requires Platinum subscription).
- Extends the delete all documents API to allow deleting documents by query.
- Enables Google Drive documents to be searched for by creator, like "created by John Doe".
Bug fixes
edit- Fixes a bug where Service Now syncs would fail when there were more than 10000 records of a single type.
- Fixes a race condition where occasionally documents can be missed from indexing.