Elasticsearch 8.6发布

Aggregations

• GeoBoundsAggregations reject sub aggregations #91073 (issue: #91072)

Allocation

• Avoid NPE when disassociateDeadNodes is executed for a node present in the desired balance #91659
• Check NodesShutdownMetadata type before assuming restart #90807

Authentication

• Fix time unit for connection request timeout of JWKs reload #92080
• Improve performance for role mapping with DNs #92074
• Improve robustness of JwkSet reloading #92081
• Support stored authentication headers prior to version 6.7 #92221

Authorization

• Make adding auth info to REST responses more robust #92168
• Security remove datemath special handling #91047

Cluster Coordination

• Fix TransportMasterNodeAction holding a CS reference needlessly #90724 (issue: #89220)
• Include last-committed data in publication #92259 (issue: #90158)
• Unsafe bootstrap memory optimization #92493

EQL

• EQL sequences: support join on multi-values #89965

Graph

• Fix potential issue with graph api’s timed out field in response #91006

Health

• Don’t account for the unassigned reason when diagnosing NO_VALID_SHARD_COPY #92416
• Fix NPE when evaluating the disk health for non-data nodes #92643
• Use https in the short URLs for the shards_availability indicator #92310

Indices APIs

• Trigger index settings providers when updating component templates #91615 (issue: #91592)

Infra/Core

• Check reserved state in Metadata.isGlobalStateEquals #92124
• Datastream unavailable exception metadata #91461
• Fix BytesRefArray on append empty BytesRef #91364
• Fix index expression options for requests with a single name or pattern #91231
• Force init of Unbox in log4j #92377 (issue: #91964)
• In file based settings, wait until security index is ready for role mappings #92173 (issue: #91939)
• Index expression exclusions never trigger “not found” #90902
• Update error states from inside the main state executor #90346 (issue: #90337)

Infra/Scripting

• Fix compile with hex literals ending with d/f #91501 (issue: #88614)

Ingest Node

• Fixing a race condition in EnrichCoordinatorProxyAction that can leave an item stuck in its queue #90688 (issue: #90598)

Machine Learning

• Copy more settings when creating DF analytics destination index #91546 (issue: #89795)
• Fix for No statistics error message #2410
• Fix for No counts available error message #2414
• Guard against input sequences that are too long for Question Answering models #91924
• Improve performance of closing files before spawning #2424
• Skip remote clusters when performing up front privileges validation for datafeeds #91895 (issue: #87832)
• Support fields with commas in data frame analytics analyzed_fields #91710 (issue: #72541)
• Validate rule filters are present on open anomaly detection api #92207

Mapping

• Consolidate field name validation when parsing mappings and documents #91328
• Fix handling empty key case in the terms aggregation #90822

Monitoring

• Fix logstash loadavg (xpack cases) #90494
• [Stack Monitoring] Update ES module mappings #90649

Network

• Clean up on exception while chunking XContent #92024
• Fix Chunked APIs sending incorrect responses to HEAD requests #92042 (issue: #92032)
• Reject connection attempts while closing #92465

SQL

• Fix NPE on logging when not tracking total hits #92425

Search

• Allow different decay values depending on the score function #91195 (issue: #78887)
• Fix timing bug with DFS profiling #92421

Snapshot/Restore

• Simplify and optimize deduplication of RepositoryData for a non-caching repository instance #91851 (issue: #89952)

Store

• Fix numOpenOutputs and modCount in ByteSizeCachingDirectory #92440 (issue: #92434)

Transform

• Skip remote clusters when performing up front privileges validation #91788

Vector Search

• Make knn search requests fully cancellable #90612

Allocation

• Deprecate state field in /_cluster/reroute response #90399
• Ensure balance threshold is at least 1 #92100

Ingest Node

• Deprecate remove_binary default of false for ingest attachment processor #90460

Mapping

• Deprecate silently ignoring type, fields, copy_to and boost in metadata field definition #90989 (issue: #35389)

Allocation

• Clear up forecasted write load and shard size from previous write index during rollovers #91590
• Forecast average shard size during rollovers #91561
• Forecast write load during rollovers #91425
• Improve shard balancing #91603
• Introduce desired-balance allocator #91343
• Limit shard realocation retries #90296
• Prevalidate node removal API #88952
• Set default cluster.routing.allocation.balance.disk_usage #91951
• Store write load in the IndexMetadata during data streams rollovers #91019
• Update the default cluster.routing.allocation.balance.disk_usage #92065
• DesiredBalance: expose it via _internal/desired_balance #91038 (issue: #90583)

Authorization

• [Fleet] Added logs-elastic_agent* read privileges to kibana_system #91701

CRUD

• Keep track of average shard write load #90768 (issue: #90102)

Geo

• Centroid aggregation for cartesian points and shapes #89216 (issue: #90156)
• Improve H3#hexRing logic and add H3#areNeighborCells method #91140
• Move SpatialUtils to geo library #88088 (issue: #86607)
• Reduce number of object allocations in H3#geoToH3 and speed up computations #91492
• Support cartesian_bounds aggregation on point and shape #91298 (issue: #90157)

ILM+SLM

• ILM: Get policy support wildcard name #89238

Infra/Core

• Handle APM global labels as affix setting #91438 (issue: #91278)
• Improve date math exclusions in expressions #90298
• Introduce a phase to use String.equals on constant strings, rather than def equality #91362 (issue: #91235)
• More actionable error for ancient indices #91243
• Operator/index templates #90143
• Operator/ingest #89735
• Transport threads and _hot_threads #90482 (issue: #90334)
• Upgrade XContent to Jackson 2.14.0 and enable Fast Double Parser #90553

Infra/Plugins

• Create placeholder plugin when loading stable plugins #90870
• Example stable plugin #90805
• Make extendedPlugins, HasNativeController and moduleName optional in plugin descriptor #90835
• Rename NamedComponent name parameter to value #91306

Infra/Scripting

• Use an explicit null check for null receivers in painless, rather than an NPE #91347 (issue: #91236)

Machine Learning

• Add a filter parameter to frequent items #91137
• Add a regex to the output of the categorize_text aggregation #90723
• Add ability to filter and sort buckets by change_point numeric values #91299
• Add api to update trained model deployment number_of_allocations #90728
• Alias timestamp to @timestamp in anomaly detection results index #90812
• Allow model_aliases to be used with Pytorch trained models #91296
• Allow overriding timestamp field to null in file structure finder #90764
• Audit a message every day the datafeed has seen no data #91774
• Low priority trained model deployments #91234 (issue: #91024)
• Provide additional information about anomaly score factors #90675

Mapping

• Don’t create IndexCaps objects when recording unmapped fields #90806 (issue: #90796)
• aggregate metric double add a max min validation #90381

Recovery

• Remove resize index settings once shards are started #90391 (issue: #90127)

Rollup

• Test downsample runtime fields and security #90593

Search

• Add LimitedOffsetsEnum to Limited offset token #86110 (issue: #86109)
• Add profiling and documentation for dfs phase #90536 (issue: #89713)
• Bulk merge field-caps responses using mapping hash #86323
• Enhance nested depth tracking when parsing queries #90425
• Expose telemetry about search usage #91528
• Return docs when using nested mappings in archive indices #90585 (issue: #90523)
• Use IndexOrDocValues query for IP range queries #90303 (issue: #83658)

Snapshot/Restore

• Increase snaphot pool max size to 10 #90282 (issue: #89608)
• Tie snapshot speed to node bandwidth settings #91021 (issue: #57023)

Store

• Allow plugins to wrap Lucene directories created by the IndexModule #91556

TLS

• Add certificate start/expiry dates to SSL Diagnostic message #89461

TSDB

• Generate index.routing_path from dynamic mapping templates #90552 (issue: #90528)
• Support malformed numbers in synthetic _source #90428
• Support synthetic _source for _doc_count field #91465
• Synthetic _source: support field in many cases #89950
• Synthetic _source: ignore_malformed for ip #90038
• Synthetic _source: support wildcard field #90196

Transform

• Add a health section to transform stats #90760
• Support aggregate_metric_double field type in transform aggregations #91045

Vector Search

• Add profiling information for knn vector queries #90200

Distributed

• Add “index” and “search” node roles with feature flag and setting #90993

EQL

• EQL samples #91312

Health

• Use chunked encoding for RestGetHealthAction #91515 (issue: #90223)
• [HealthAPI] Use the RestCancellableNodeClient infrastructure #91587

Machine Learning

• Make categorize_text aggregation GA #88600

Vector Search

• Add fielddata and scripting support for byte-sized vectors #91184
• Add support for indexing byte-sized knn vectors #90774

Infra/Core

• Revert “Remove ImmutableOpenMap from snapshot services” #90287

Infra/Logging

• Upgrade to log4j 2.19.0 #90589 (issue: #90584)

Network

• Upgrade to Netty 4.1.82.Final #90604
• Upgrade to Netty 4.1.84 #91271

Snapshot/Restore

• Upgrade GCS SDK to 2.13.1 #92327

转自 Elasticsearch version 8.6.0 | Elasticsearch Guide [8.6] | Elastic