7.14.0 release notes

Upgrading to Enterprise Search 7.14.0? See Upgrading & migrating.

New interface, new possibilities. Enterprise Search is permanently moving to Elastic’s main management interface - known as Kibana - with the 8.0 release. Try the new interface in Kibana beginning in 7.14. Learn more about the exciting new possibilities, and important migration timelines. See User interfaces.
New user management UI.
Speeds up migrations from older versions, especially for larger deployments with a lot of indexed documents.
New config options allow retrying the connection to Kibana.
Deploy using aarch64 .rpm packages and Docker images.
Deploy using arm64 .deb packages.

Prior to 7.14.0, Enterprise Search had three auth modes: standard, native, and SAML. Auth modes are deprecated in Enterprise Search 7.14.0 and will be removed in 8.0.0.
- However, Enterprise Search will use your current configuration to migrate your existing user data. When upgrading from Enterprise Search 7.13 and earlier, do not change any user/auth configuration settings before upgrading to 7.14.0.
- Refer to the following documentation for details: Upgrade from Enterprise Search 7.13 and earlier and Users and access.
Startup now requires that the version of Elasticsearch is compatible with the major and minor version of Enterprise Search.
If there are no auth sources configured, but kibana.external_url is present, users will only be able to access the Enterprise Search UI from Kibana.

Reduces the number of Elasticsearch requests made by the /api/ent/v1/internal/stats API.
Supports non-latin (UTF-8) characters in the URL path.

Web crawler supports excluding and including specific parts of the <body> content using data-elastic-exclude and data-elastic-include attributes.
Web crawler validates domains and warns of common problems before crawling.
Web crawler provides metrics and diagnostics APIs.
Web crawler URL debugger API.
Precision slider control to adjust relevance and precision, available in the Enterprise Search Kibana plugin UI.

Ensures that search boost factors are greater or equal to zero.
- User Interface does not allow choosing negative values for boost factors.
Fixes a bug where App Search API keys with read access and full engine access were inadvertently also able to create engines. API keys now require write access to be able to create engines.
Fixes a bug where a dev user with scoped engine access could view all credentials via the /api/as/v1/credentials endpoint.
Fixes a bug where a dev user without full engine access could programmatically create API keys with full engine access.
When processing a sitemap with thousands of URLs, canceling a crawl will now finish in a timely manner.

bulk_create endpoint now supports a last_updated field when indexing and updating documents.
Adds an API that allows a content source to be refreshed ad-hoc with POST /api/ws/v1/sources/<source_id>/refresh.
Adds a GET /api/ws/v1/whoami API endpoint to get details about the currently authenticated user.
Configure more optional display settings for custom content sources.
Upload an icon for a custom source.

Documents pushed via custom source APIs now have an updated_at field to track the time at which they were indexed. Integrations built using the custom source APIs that were already specifying a field named updated_at will need to choose a new field name to avoid a collision.

Fixes a bug where Google Drive incremental syncs did not pick up changes made outside the connecting user’s drive by other users.
Fixes a bug with duplicate content sources displayed to the user when user belongs to multiple groups that have access to same content sources.
Fixes a bug where some Confluence Cloud errors could not be logged and would fail connector jobs.
Addresses a change from the Confluence Cloud OAuth API introduced on 2021-06-24 where support for <base_url>/wiki with OAuth bearer tokens was removed.
Fixes a bug where the Dropbox connector stops working after access token expiration.