« Release notes Elasticsearch version 8.19.21 »
Elastic Docs ›Elasticsearch Guide [8.19] ›Release notes

Elasticsearch version 8.19.22

edit
IMPORTANT: This documentation is no longer updated. Refer to Elastic's version policy and the latest documentation.

Elasticsearch version 8.19.22

edit

Coming in 8.19.22.

Also see Breaking changes in 8.19.

Bug fixes

edit
Analysis
  • Add option to limit the number of characters in analyze filters #158713
  • Bound min_hash filter parameters to prevent out-of-memory errors #154480
  • Cap min_hash filter parameters #158211
Authentication
  • Fix stack overflow when authorizing LDAP users with large group membership #158564
  • Scope role mapping metadata to gateway and API #158862
Authorization
  • Reject allow_restricted_indices=true consistently in roles granted by users with the manage_roles privilege #158831
ES|QL
  • Fix ST_DISTANCE with both arguments from doc values #157983 (issue: #149814)
  • Fix bulk inference requests being left waiting forever #158313 (issue: #156135)
  • Reject 2^63 in QL and ES|QL double-to-long conversion #157555 (issue: #157551)
  • When reversing strings, treat carriage return line feed (CRLF) as a single control character, in line with Unicode guidelines #158745
Highlighting
  • Allow max_analyzed_offset to be set to -1 to default to index.highlight.max_analyzed_offset #157381 (issue: #156122)
Machine Learning
  • Extend anomaly detection results template for reindexed indices and generalize healing beyond job_id mapping #153755 (issue: #147686)
  • Fix NPE when preview datafeed checks date_nanos for an unmapped time field #144909 (issue: #144888)
  • Fix model.failed_category_count in _cat/ml/anomaly_detectors reporting the dead category count #158186
  • Limit concurrent text structure analyses to the allocated processor count #158935 (issue: #158927)
  • Reject blank datafeed ID in stop datafeed request #157929
  • Reject blank job ID on close request #158020
  • Prevent STOPPING trained model routes with no local deployment process from blocking ML node shutdown during node drain #158820 (issue: #158818)
  • Treat ML node exhaustion as 4xx #155310
SQL
  • Fix result size limit check on INSERT #157752
Search
  • Keep a cancelled search from failing others through the request cache #157457 (issue: #157302)
  • Release fetch-source request breaker bytes #158734 (issue: #158634)
  • Prevent scripted fields from causing out-of-memory errors #152718 (issue: #120459)
Snapshot/Restore
  • Fix interrupt deadlock in CancellableRateLimitedFluxIterator #158936
Suggesters
  • Cap phrase suggester token_limit and bound candidate scoring work #158434

Enhancements

edit
Search
  • Bring _msearch/template to parity with _msearch: memory bounds, cancellation, and metrics #157642

Upgrades

edit
Infra/Core
Network
  • Upgrade Netty to 4.1.137.Final #157348
Security
« Release notes Elasticsearch version 8.19.21 »