Elasticsearch 8.5发布

CRUD

• Reject unknown bulk actions #89450

Aggregations

• Fix auto_date_histogram > ip_range #90317 (issue: #90121)
• Fail when rebuilding scorer in breadth_first mode and query context has changed #89993 (issue: #37650)
• Fix merging with empty results #86939 (issue: #84622)
• Fix partial reduce bug in ip_prefix #89734 (issue: #89686)
• Remove unexpected meta parameter in aggregation response #89467 (issue: #89455)
• Consistently allow sorting top_metrics aggregations whose size is greater than 1 #89974 (issue: #86663)
• Check parent circuit breaker when allocating an empty bucket #89568 (issue: #80789)

Authorization

• Remove magic string for the "__empty" role #89766

CCR

• Order follower backing indices for data streams by original name #90850 (issue: #90820)

Cluster Coordination

• Capture deprecation warnings in batched master tasks #85525 (issue: #85506)
• Check circuit breaker before sending join request #89318 (issue: #85003)

Distributed

• Fork TransportClusterStateAction to MANAGEMENT thread pool #90996

Engine

• Fix "path.conf'" typo in Security.java #89248 (issue: #89327)

Geo

• Buffer H3Polygon2D bounding boxes to avoid edge precision issues #89196 (issues: #89868, #87391)
• Fix date histogram range edge case #88957
• Format runtime geo_points #85449 (issue: #85245)
• Generate error if scripts load _source with synthetic _source enabled #88334

Graph

• Fix race condition in timeout #88946 (issue: #55396)

Health

• Modify disk indicator details to provide an overview of the disk space health #90189
• Fix disk indicator impacts and diagnosis #90262
• Fix the details calculation of the disk indicator #90869
• Report impact and diagnosis data nodes without disk space and no blocked indices #90772 (issue: #90442)
• Update minimum version for health node reporting to 8.5 #90365 (issue: #90359)

Indices APIs

• Avoid capturing per-task RolloverResult #90626 (issue: #90620)
• Fix shard splitting for nested #89351 (issue: #88109)

Infra/Core

• Fix allow_no_indices request option in special cases #89622
• Fix repeated error save loops in File Settings Service #90271 (issue: #90222)
• Fix date rounding for date math parsing #90458 (issue: #90187)
• Fix disabling APM tracing for CancellableTask in TrainedModelAssignmentNodeService #90972 (issue: #89850)
• Support camel case dates on 7.x indices #88914 (issue: #84199)

Infra/Scripting

• Fix true/false accumulation bug in boolean source fallback #90895

Ingest Node

• Set the enrich maintenance cluster lifecycle listener only once #90486

Machine Learning

• Require correct tier processors when multiple AZs are present #90903
• Return 408 instead of 500 when open/start APIs time out #89775 (issue: #89585)

Mapping

• Fix duplication bug for source fallback in numeric types #89352
• Include runtime fields in total fields count #89251 (issue: #88265)
• Fix aggregate_metric_double multi-values exception #90290
• Validate field names when subobjects are disabled #90950

Monitoring

• Add fields to fix Logstash cgroup graphs #90493

Network

• Fix RecyclerBytesStreamOutput allocating unlimited heap for some capacities #90632

Recovery

• Fix overcounting recovered bytes after network disconnect #90477 (issue: #90441)

Search

• Add support for predefined char class regexp on wildcard fields #90064
• Deduplicate fetching doc-values fields #89094
• Don’t shortcut the total hit count for text fields #90341 (issue: #89760)
• Safeguard RegExp use against StackOverflowError #84624 (issue: #82923)
• Use MB rather than GB to calculate max boolean clauses #90309 (issue: #86136)

Snapshot/Restore

• Fix incorrect failed shards count in APIs for current snapshots #89534
• Fix over-allocation of mounted indices on a cold/frozen node #86331
• Fix quadratic complexity in SnapshotStatus serialization #90795
• Fork building snapshot status response off of transport thread #90651
• Make sure listener is resolved when file queue is cleared #89929
• Re-register a corrupt repository to unblock it #89719 (issue: #89130)
• Reject unknown request body fields in mount API #88987 (issue: #75982)

TSDS

• Fix segment stats in TSDS #89754 (issue: #89609)
• Fix extra fields in GET request for synthetic _source #89778
• Fix scaled_float rounding for synthetic _source #88916 (issue: #88854)

Transform

• Don’t fail a transform on a ClusterBlockException, this may be due to ILM closing an index #90396 (issue: #89802)
• Fix NPE in transform scheduling #90347 (issues: #90356, #88203, #90301, #90255)
• Improve error handling in state persistence #88910 (issue: #88905)
• Return 408 instead of 500 when the start API times out #89774

Vector Search

• Fix bug for kNN with filtered aliases #89621

Watcher

• Allow xpack.notification.email.account.domain_allowlist to be set dynamically #90426 (issue: #89913)
• Handling timeout exceptions on watcher startup #90421 (issue: #44981)

Infra/Plugins

• Deprecate network plugins #88924
• Deprecate overriding DiscoveryPlugin internals #88925

Authentication

• Add more accurate error message for LDAP user modes #89492

Authorization

• Add indices permissions to Enterprise Search service account #89869
• Add information of resolved roles in denial messages #89680

Autoscaling

• Centralize the concept of processors configuration #89662

Cluster Coordination

• Preemptively compute RoutingNodes and the indices lookup during publication #89005
• Preemptively initialize routing nodes and indices lookup on all node types #89032

Distributed

• Batch index delete cluster state updates #90033 (issue: #90022)
• Increase the minimum size of the management pool to 2 #90193

Health

• Add IDs to health API diagnoses and impacts #90072
• Add a check to the master stability health API when there is no master and the current node is not master eligible #89219
• Add logic to master_is_stable indicator to check for discovery problems #88020
• Poll for cluster diagnostics information #89014
• Update SLM health diagnosis message to include unhealthy policy details #89138

Highlighting

• Improve efficiency of BoundedBreakIteratorScanner fragmentation algorithm #89041 (issues: #73569, #73785)

ILM+SLM

• Add validations for the downsampling ILM action #90295
• Ensure that ILM does not roll over empty indices #89557 (issue: #86203)
• Reuse informational message in lifecycle step #89419
• Move log-related logic into log block in IndexLifecycleRunner #89292

Infra/Core

• Add reserved /_snapshot/repo file based settings #89601
• Add upgrade_status attributes to Fleet Agents #89845
• Add support for /_autoscaling/policy for file based settings #89708
• Add support for /_security/role_mapping for file based settings #89667
• Add support for support for /_slm/policy in file based settings #89567
• Retry file watch registration #90537 (issue: #89500)

Infra/Node Lifecycle

• Distinguish no shutdowns case in NodeShutdownAllocationDecider #89851 (issue: #89823)

Infra/Plugins

• Add deprecation message for deprecated plugin APIs #88961
• Register stable plugins in ActionModule #90067
• Load plugin named components #89969

Infra/Scripting

• Initial code to support binary expression scripts #89895
• Protect _source inside update scripts #88733
• Reindex and UpdateByQuery metadata #88665
• Add write Field API NestedDocument support #90021
• Add write Field API path manipulation #89889
• Add write Field API with basic path resolution #89738
• Add write Fields API for reindex, update, and update by query #90145

Infra/Settings

• Introduce max headroom for disk watermark stages #88639 (issue: #81406)

License

• License check for user profile collaboration feature #89990

Machine Learning

• Add measure of non-cache hit inference count #90464
• Add new text_similarity nlp task #88439
• Add new trained model deployment cache clear API #89074
• Add processor autoscaling decider #89645
• Distribute trained model allocations across availability zones #89822
• Use a bitset for deduplication of frequent items #88943
• Optimize frequent items transaction lookup #89062
• Release native inference functionality as beta #90418
• Return 408 when the start deployment API times out #89612
• Skip renormalization after calling the node shutdown API #89347
• Compute outlier feature influence via the Gateaux derivative to improve attribution for high dimension vectors #2256
• Improve classification and regression model train runtimes for data sets with many numeric features #2380, #2388, #2390, #2401
• Increase the limit on the maximum number of classes to 100 for training classification models #2395 (issue: #2246)

Mapping

• Add synthetic_source support to aggregate_metric_double fields #88909
• Add source fallback for keyword fields using operation #88735
• Add source fallback support for match_only_text mapped type #89473
• Add source fallback support for date and date_nanos mapped types #89440
• Add source fallback support for unsigned long mapped type #89349
• Add support for source fallback with scaled float field type #89053
• Add support for source fallback with the boolean field type #89052
• Add text field support in the Painless scripting fields API #89396
• Clarify that fielddata is not supported for text fields error message #89770 (issue: #89485)
• Add new mappings for Fleet Agent last_checkin_message and components fields #89599
• Support source fallback for byte, short, and long fields #88954
• Support source fallback for double, float, and half_float field types #89010

Network

• Use chunked REST serialization for large REST responses #88311

Recovery

• Disable recovering from snapshots in searchable snapshots #86388

SQL

• Implement DATE_FORMAT function #88388 (issue: #55065)
• Set track_total_hits to false when not needed #89106 (issue: #88764)

Search

• Enable BloomFilter for _id of non-datastream indices #88409
• In the field capabilities API, renew support for fields in the request body #88972 (issue: #86875)

Security

• Add usage stats report for user profiles #90123
• Implement grace period for user profile activation #89566
• Return limited-by role descriptors in Get/QueryApiKey response #89273
• Add option to return profile uid in GetUser response #89570
• Return 400 error for GetUserPrivileges call with API keys #89333
• Show assigned role descriptors in Get/QueryApiKey response #89166
• Add detailed errors in hasPrivileges response #89224
• Add support for multiple UIDs to the GetProfile API #89023

Snapshot/Restore

• Add support for comparing SnapshotsInProgress #89619 (issue: #88732)
• Prioritize shard snapshot tasks over file snapshot tasks and limit the number of the concurrently running snapshot tasks #88209 (issue: #83408)

Stats

• Introduce node mappings stats #89807

TSDS

• Support match_only_text for synthetic _source #89516
• Support histogram field for synthetic _source #89833
• Support version field type for synthetic _source #89706
• Build _id without reparsing #88789
• Return metric fields in the field caps API #88695

Transform

• Add an unattended mode setting to transform #89212

Authorization

• Introduce the new read_security cluster privilege #89790 (issue: #89245)

Health

• Enable the health node and the disk health indicator #90085 (issue: #84811)

Infra/Core

• Provide tracing implementation using OpenTelemetry and APM Java agent #88443 (issue: #84369)

Infra/Plugins

• Add the stable Plugin API module and analysis interfaces #88775

Machine Learning

• Make bucket_correlation aggregation generally available #88655
• Make bucket_count_ks_test aggregation generally available #88657

Security

• Support bulk updates of API keys #88856

TSDS

• Add a TSID global ordinal to TimeSeriesIndexSearcher #90035
• Release time series data stream functionality #90116 (issue: #74660)
• Add synthetic _source support for the ignore_above parameter on keyword fields #89466

Vector Search

• Add synthetic _source support for dense_vector #89840

Infra/Scripting

• Fix fields API caching regression #90017

Client

• Upgrade Apache Commons Logging to 1.2 #85745 (issue: #40305)

Packaging

• Upgrade bundled JDK to Java 19 #90571

转自 Elasticsearch version 8.5.0 | Elasticsearch Guide [8.5] | Elastic