8.15.0 release notes
edit8.15.0 release notes
editUpgrading to Enterprise Search 8.15.0? See Upgrading and migrating.
Known issues
edit- Enterprise Search: Enabling [APM for Enterprise Search](https://www.elastic.co/guide/en/enterprise-search/current/monitoring.html#apm-for-enterprise-search) does not work in this version and will return the following error:
[ElasticAPM] Couldn't establish connection to APM Server: "#<Java::JavaLang::NoSuchMethodError: 'org.bouncycastle.asn1.DERIA5String org.bouncycastle.asn1.DERIA5String.getInstance(java.lang.Object)'>" [ElasticAPM] Connection error: #<Errno::EPIPE: Broken pipe - No message available>
Users will observe the above errors in the Elastic Cloud Enterprise Search logs because APM integration is enabled by default on Elastic Cloud.
New features
edit- The Sharepoint Server connector is now available as a native connector (managed service) on Elastic Cloud.
-
The following connectors now support document-level security (DLS):
- Outlook
- Sharepoint Server
- We added incremental sync support for the Notion connector. (See PR #2547)
-
We introduced a feature flag
elasticsearch.feature_use_connectors_api
to enable using Elasticsearch Connector APIs (beta) to manage connectors via system indices instead of direct dot-prefixed index access. In the future, this flag will be enabled by default. (See PR #2561)
Enhancements
edit-
API keys generated in Kibana for connectors will now have the new
manage_connector
andmonitor_connector
privileges. -
We added a new optional field in advanced sync rules for the PostgreSQL connector named
id_columns
. Use this field to ingest tables which do not have a primary key. Include the names of unique fields so that the connector can use them to generate unique IDs for documents.(See PR #2681) -
We added a new
Custom_Fields
field in the Jira connector to display custom fields separately. (See PR #2505) - We added the ability to sync labels from pages in the Confluence connector. (See PR #2549)
Bug fixes
edit- Fixed a bug where the connector framework might receive an interrupt, but not terminate all running sync job tasks correctly. (See PR #2661)
- We removed unnecessary logging from the S3 connector. (See PR #2656)
- Fixed a bug where sync jobs could be permanently stuck in "pending" (until manually canceled) because of an unfortunately timed Elasticsearch exception. Now such exceptions will cause jobs to fail and provide actionable error messaging. (See PR #2606)
- Fixed connector scheduling not working for "Every Week" and "Every Year" statements. (See PR #2594)
- Fixed a bug where the incremental sync scheduling dropdown didn’t include the "Every Minute" option. (See PR #189250)
- Fixed a bug where the Confluence connector would fail to sync complete documents when using advanced sync rules. (See PR #2602)