OpenSearch

• Add new API to modify data stream backing indices and multivalue_doc_count aggregation for counting documents with multiple values
• Add HTTP/3 client support, gRPC support for Point-in-Time operations, and cross-cluster streaming via Arrow Flight
• Improve query performance with coordinator-side field domain pruning and add security enhancements including Java deserialization filtering
• Fix multiple bugs including snapshot listing failures, S3 encryption issues, and stack overflow vulnerabilities in deserialization

• Add dynamic properties support for pattern-based field definitions and pluggable data format engine with multi-format indexing capabilities
• Implement WritableWarm tiered storage with merge support, prefetch settings, and comprehensive metrics for warm node management
• Expand PPL analytics engine with 100+ new scalar functions, window functions, and commands (append, dedup, eventstats, parse, rex, spath, top, rare)
• Add distributed aggregation, joins, and native Arrow transport for analytics engine with improved memory management and query cancellation

• Pull-based ingestion is now a public API (no longer experimental), with new warmup phase and field mapping support
• Multiple performance improvements including adaptive shard selection for bulk writes, intra-segment metric aggregations, and terms aggregation fixes
• New bitmap64 query support and enhanced search capabilities with workload management and scroll API support
• Various bug fixes for geo_point fields, field_caps, segment replication, and TLS certificate hot-reload

• Support for HTTP/3 on the server side and Alt-Svc header to advertise HTTP/3 availability
• New circuit breaker for gRPC transport to prevent out-of-memory errors
• Index-level encryption support for snapshots and remote-store
• Concurrent translog recovery to accelerate segment replication primary promotion

• Added support for ForkJoinPool type and pluggable gRPC interceptors with explicit ordering
• New cluster setting search.query.max_query_string_length_monitor_only for query monitoring
• Performance improvements for percentiles aggregation, matrix_stats aggregation, and terms query creation for keyword fields
• Multiple bug fixes including wildcard query handling, GRPC bulk operations, and snapshot restoration issues

• Fixed sub-aggregator casting issue when using search profiling with Star Tree
• Fixed backward compatibility issue with @timestamp upgrades
• Fixed IndexOutOfBoundsException in terms aggregations with include/exclude filters
• Fixed various other issues including core patching and S3 repository settings

• Fixed an issue when updating core with a patch number other than 0
• Fixed sub-aggregator casting for Star Tree searches with profiling enabled
• Fixed backward compatibility issue with @timestamp upgrade by adding version check on skip_list parameter

• Added support for temporal routing processors, query rewriting infrastructure, and streaming aggregations with new aggregator types
• Expanded profiling capabilities for fetch phase including inner hits and top hits aggregations
• Implemented multiple gRPC query types (MatchPhrase, Boolean, Script, Exists, Regexp, Wildcard, Geo, Range, Nested) for improved transport efficiency
• Added rule-based auto-tagging with security attributes and multiple attribute support

• Fixed hangs in bulk requests when an index is deleted during the primary phase
• Fixed case-insensitive and escaped queries on wildcard fields
• Fixed array index out of bounds exceptions with wildcard fields and aggregations
• Fixed circuit breaker logic to prevent negative estimated limits and fielddata stats

• Fixed segment replication failures that could occur during rolling restarts
• Fixed case sensitivity handling in XContentMapValues.filter
• Updated dependencies including Netty, log4j, and shadow-gradle-plugin

• New cluster setting search.query.max_query_string_length to control query string length limits
• Fixed out-of-memory issue caused by large numbers of shard results being buffered
• Fixed IndexOutOfBoundsException in terms aggregations with include/exclude on non-existent prefixes
• Multiple dependency updates including Lucene, Netty, Reactor, and security libraries

No changelog text published for this release.