elasticsearch/docs/reference/release-notes/8.15.0.asciidoc

[[release-notes-8.15.0]]
== {es} version 8.15.0

Also see <<breaking-changes-8.15,Breaking changes in 8.15>>.


[[known-issues-8.15.0]]
[float]
=== Known issues
* The `pytorch_inference` process used to run Machine Learning models can consume large amounts of memory.
In environments where the available memory is limited, the OS Out of Memory Killer will kill the `pytorch_inference`
process to reclaim memory. This can cause inference requests to fail.
Elasticsearch will automatically restart the `pytorch_inference` process
after it is killed up to four times in 24 hours. (issue: {es-issue}110530[#110530])

* Pipeline aggregations under `time_series` and `categorize_text` aggregations are never
returned (issue: {es-issue}111679[#111679])

* Elasticsearch will not start on Windows machines if
[`bootstrap.memory_lock` is set to `true`](https://www.elastic.co/guide/en/elasticsearch/reference/current/setup-configuration-memory.html#bootstrap-memory_lock).
Either downgrade to an earlier version, upgrade to 8.15.1, or else follow the
recommendation in the manual to entirely disable swap instead of using the
memory lock feature (issue: {es-issue}111847[#111847])

* The `took` field of the response to the <<docs-bulk>> API is incorrect and may be rather large. Clients which
<<api-conventions-number-values,incorrectly>> assume that this value will be within a particular range (e.g. that it fits into a 32-bit
signed integer) may encounter errors (issue: {es-issue}111854[#111854])

* Elasticsearch will not start if custom role mappings are configured using the
`xpack.security.authc.realms.*.files.role_mapping` configuration option. As a workaround, custom role mappings
can be configured using the https://www.elastic.co/guide/en/elasticsearch/reference/current/security-api-put-role-mapping.html[REST API] (issue: {es-issue}112503[#112503])

* ES|QL queries can lead to node crashes due to Out Of Memory errors when:
** Multiple indices match the query pattern
** These indices have many conflicting field mappings
** Many of those fields are included in the request
These issues deplete heap memory, increasing the likelihood of OOM errors. (issue: {es-issue}111964[#111964], {es-issue}111358[#111358]).
In Kibana, you might indirectly execute these queries when using Discover, or adding a Field Statistics panel to a dashboard.
+
To work around this issue, you have a number of options:
** Downgrade to an earlier version
** Upgrade to 8.15.2 upon release
** Follow the instructions to
<<esql-kibana-enable,disable ES|QL queries in {kib}>>
** Change the default data view in Discover to a smaller set of indices and/or one with fewer mapping conflicts.

[[breaking-8.15.0]]
[float]
=== Breaking changes

Cluster Coordination::
* Interpret `?timeout=-1` as infinite ack timeout {es-pull}107675[#107675]

Inference API::
* Replace `model_id` with `inference_id` in GET inference API {es-pull}111366[#111366]

Rollup::
* Disallow new rollup jobs in clusters with no rollup usage {es-pull}108624[#108624] (issue: {es-issue}108381[#108381])

Search::
* Change `skip_unavailable` remote cluster setting default value to true {es-pull}105792[#105792]

[[bug-8.15.0]]
[float]
=== Bug fixes

Aggregations::
* Don't sample calls to `ReduceContext#consumeBucketsAndMaybeBreak` ins `InternalDateHistogram` and `InternalHistogram` during reduction {es-pull}110186[#110186]
* Fix `ClassCastException` in Significant Terms {es-pull}108429[#108429] (issue: {es-issue}108427[#108427])
* Run terms concurrently when cardinality is only lower than shard size {es-pull}110369[#110369] (issue: {es-issue}105505[#105505])

Allocation::
* Fix misc trappy allocation API timeouts {es-pull}109241[#109241]
* Fix trappy timeout in allocation explain API {es-pull}109240[#109240]

Analysis::
* Correct positioning for unique token filter {es-pull}109395[#109395]

Authentication::
* Add comma before charset parameter in WWW-Authenticate response header {es-pull}110906[#110906]
* Avoid NPE if `users_roles` file does not exist {es-pull}109606[#109606]
* Improve security-crypto threadpool overflow handling {es-pull}111369[#111369]

Authorization::
* Fix trailing slash in `security.put_privileges` specification {es-pull}110177[#110177]
* Fixes cluster state-based role mappings not recovered from disk {es-pull}109167[#109167]
* Handle unmatching remote cluster wildcards properly for `IndicesRequest.SingleIndexNoWildcards` requests {es-pull}109185[#109185]

Autoscaling::
* Expose `?master_timeout` in autoscaling APIs {es-pull}108759[#108759]

CRUD::
* Update checkpoints after post-replication actions, even on failure {es-pull}109908[#109908]

Cluster Coordination::
* Deserialize publish requests on generic thread-pool {es-pull}108814[#108814] (issue: {es-issue}106352[#106352])
* Fail cluster state API if blocked {es-pull}109618[#109618] (issue: {es-issue}107503[#107503])
* Use `scheduleUnlessShuttingDown` in `LeaderChecker` {es-pull}108643[#108643] (issue: {es-issue}108642[#108642])

Data streams::
* Apm-data: set concrete values for `metricset.interval` {es-pull}109043[#109043]
* Ecs@mappings: reduce scope for `ecs_geo_point` {es-pull}108349[#108349] (issue: {es-issue}108338[#108338])
* Include component templates in retention validaiton {es-pull}109779[#109779]

Distributed::
* Associate restore snapshot task to parent mount task {es-pull}108705[#108705] (issue: {es-issue}105830[#105830])
* Don't detect `PlainActionFuture` deadlock on concurrent complete {es-pull}110361[#110361] (issues: {es-issue}110181[#110181], {es-issue}110360[#110360])
* Handle nullable `DocsStats` and `StoresStats` {es-pull}109196[#109196]

Downsampling::
* Support flattened fields and multi-fields as dimensions in downsampling {es-pull}110066[#110066] (issue: {es-issue}99297[#99297])

ES|QL::
* ESQL: Change "substring" function to not return null on empty string {es-pull}109174[#109174]
* ESQL: Fix Join references {es-pull}109989[#109989]
* ESQL: Fix LOOKUP attribute shadowing {es-pull}109807[#109807] (issue: {es-issue}109392[#109392])
* ESQL: Fix Max doubles bug with negatives and add tests for Max and Min {es-pull}110586[#110586]
* ESQL: Fix `IpPrefix` function not handling correctly `ByteRefs` {es-pull}109205[#109205] (issue: {es-issue}109198[#109198])
* ESQL: Fix equals `hashCode` for functions {es-pull}107947[#107947] (issue: {es-issue}104393[#104393])
* ESQL: Fix variable shadowing when pushing down past Project {es-pull}108360[#108360] (issue: {es-issue}108008[#108008])
* ESQL: Validate unique plan attribute names {es-pull}110488[#110488] (issue: {es-issue}110541[#110541])
* ESQL: change from quoting from backtick to quote {es-pull}108395[#108395]
* ESQL: make named params objects truly per request {es-pull}110046[#110046] (issue: {es-issue}110028[#110028])
* ES|QL: Fix DISSECT that overwrites input {es-pull}110201[#110201] (issue: {es-issue}110184[#110184])
* ES|QL: limit query depth to 500 levels {es-pull}108089[#108089] (issue: {es-issue}107752[#107752])
* ES|QL: reduce max expression depth to 400 {es-pull}111186[#111186] (issue: {es-issue}109846[#109846])
* Fix ST_DISTANCE Lucene push-down for complex predicates {es-pull}110391[#110391] (issue: {es-issue}110349[#110349])
* Fix `ClassCastException` with MV_EXPAND on missing field {es-pull}110096[#110096] (issue: {es-issue}109974[#109974])
* Fix bug in union-types with type-casting in grouping key of STATS {es-pull}110476[#110476] (issues: {es-issue}109922[#109922], {es-issue}110477[#110477])
* Fix for union-types for multiple columns with the same name {es-pull}110793[#110793] (issues: {es-issue}110490[#110490], {es-issue}109916[#109916])
* [ESQL] Count_distinct(_source) should return a 400 {es-pull}110824[#110824]
* [ESQL] Fix parsing of large magnitude negative numbers {es-pull}110665[#110665] (issue: {es-issue}104323[#104323])
* [ESQL] Migrate `SimplifyComparisonArithmetics` optimization {es-pull}109256[#109256] (issues: {es-issue}108388[#108388], {es-issue}108743[#108743])

Engine::
* Async close of `IndexShard` {es-pull}108145[#108145]

Highlighting::
* Fix issue with returning incomplete fragment for plain highlighter {es-pull}110707[#110707]

ILM+SLM::
* Allow `read_slm` to call GET /_slm/status {es-pull}108333[#108333]

Indices APIs::
* Create a new `NodeRequest` for every `NodesDataTiersUsageTransport` use {es-pull}108379[#108379]

Infra/Core::
* Add a cluster listener to fix missing node features after upgrading from a version prior to 8.13 {es-pull}110710[#110710] (issue: {es-issue}109254[#109254])
* Add bounds checking to parsing ISO8601 timezone offset values {es-pull}108672[#108672]
* Fix native preallocate to actually run {es-pull}110851[#110851]
* Ignore additional cpu.stat fields {es-pull}108019[#108019] (issue: {es-issue}107983[#107983])
* Specify parse index when error occurs on multiple datetime parses {es-pull}108607[#108607]

Infra/Metrics::
* Provide document size reporter with `MapperService` {es-pull}109794[#109794]

Infra/Node Lifecycle::
* Expose `?master_timeout` on get-shutdown API {es-pull}108886[#108886]
* Fix serialization of put-shutdown request {es-pull}107862[#107862] (issue: {es-issue}107857[#107857])
* Support wait indefinitely for search tasks to complete on node shutdown {es-pull}107426[#107426]

Infra/REST API::
* Add some missing timeout params to REST API specs {es-pull}108761[#108761]
* Consider `error_trace` supported by all endpoints {es-pull}109613[#109613] (issue: {es-issue}109612[#109612])

Ingest Node::
* Fix Dissect with leading non-ascii characters {es-pull}111184[#111184]
* Fix enrich policy runner exception handling on empty segments response {es-pull}111290[#111290]
* GeoIP tasks should wait longer for master {es-pull}108410[#108410]
* Removing the use of Stream::peek from `GeoIpDownloader::cleanDatabases` {es-pull}110666[#110666]
* Simulate should succeed if `ignore_missing_pipeline` {es-pull}108106[#108106] (issue: {es-issue}107314[#107314])

Machine Learning::
* Allow deletion of the ELSER inference service when reference in ingest {es-pull}108146[#108146]
* Avoid `InferenceRunner` deadlock {es-pull}109551[#109551]
* Correctly handle duplicate model ids for the `_cat` trained models api and usage statistics {es-pull}109126[#109126]
* Do not use global ordinals strategy if the leaf reader context cannot be obtained {es-pull}108459[#108459]
* Fix NPE in trained model assignment updater {es-pull}108942[#108942]
* Fix serialising inference delete response {es-pull}109384[#109384]
* Fix "stack use after scope" memory error {ml-pull}2673[#2673]
* Fix trailing slash in `ml.get_categories` specification {es-pull}110146[#110146]
* Handle any exception thrown by inference {ml-pull}2680[#2680]
* Increase response size limit for batched requests {es-pull}110112[#110112]
* Offload request to generic threadpool {es-pull}109104[#109104] (issue: {es-issue}109100[#109100])
* Propagate accurate deployment timeout {es-pull}109534[#109534] (issue: {es-issue}109407[#109407])
* Refactor TextEmbeddingResults to use primitives rather than objects {es-pull}108161[#108161]
* Require question to be non-null in `QuestionAnsweringConfig` {es-pull}107972[#107972]
* Start Trained Model Deployment API request query params now override body params {es-pull}109487[#109487]
* Suppress deprecation warnings from ingest pipelines when deleting trained model {es-pull}108679[#108679] (issue: {es-issue}105004[#105004])
* Use default translog durability on AD results index {es-pull}108999[#108999]
* Use the multi node routing action for internal inference services {es-pull}109358[#109358]
* [Inference API] Extract optional long instead of integer in `RateLimitSettings#of` {es-pull}108602[#108602]
* [Inference API] Fix serialization for inference delete endpoint response {es-pull}110431[#110431]
* [Inference API] Replace `model_id` with `inference_id` in inference API except when stored {es-pull}111366[#111366]

Mapping::
* Fix off by one error when handling null values in range fields {es-pull}107977[#107977] (issue: {es-issue}107282[#107282])
* Limit number of synonym rules that can be created {es-pull}109981[#109981] (issue: {es-issue}108785[#108785])
* Propagate mapper builder context flags across nested mapper builder context creation {es-pull}109963[#109963]
* `DenseVectorFieldMapper` fixed typo {es-pull}108065[#108065]

Network::
* Use proper executor for failing requests when connection closes {es-pull}109236[#109236] (issue: {es-issue}109225[#109225])
* `NoSuchRemoteClusterException` should not be thrown when a remote is configured {es-pull}107435[#107435] (issue: {es-issue}107381[#107381])

Packaging::
* Adding override for lintian false positive on `libvec.so` {es-pull}108521[#108521] (issue: {es-issue}108514[#108514])

Ranking::
* Fix score count validation in reranker response {es-pull}111424[#111424] (issue: {es-issue}111202[#111202])

Rollup::
* Fix trailing slash in two rollup specifications {es-pull}110176[#110176]

Search::
* Adding score from `RankDoc` to `SearchHit` {es-pull}108870[#108870]
* Better handling of multiple rescorers clauses with LTR {es-pull}109071[#109071]
* Correct query profiling for conjunctions {es-pull}108122[#108122] (issue: {es-issue}108116[#108116])
* Fix `DecayFunctions'` `toString` {es-pull}107415[#107415] (issue: {es-issue}100870[#100870])
* Fix leak in collapsing search results {es-pull}110927[#110927]
* Fork freeing search/scroll contexts to GENERIC pool {es-pull}109481[#109481]

Security::
* Add permission to secure access to certain config files {es-pull}107827[#107827]
* Add permission to secure access to certain config files specified by settings {es-pull}108895[#108895]
* Fix trappy timeouts in security settings APIs {es-pull}109233[#109233]

Snapshot/Restore::
* Stricter failure handling in multi-repo get-snapshots request handling {es-pull}107191[#107191]

TSDB::
* Sort time series indices by time range in `GetDataStreams` API {es-pull}107967[#107967] (issue: {es-issue}102088[#102088])

Transform::
* Always pick the user `maxPageSize` value {es-pull}109876[#109876] (issue: {es-issue}109844[#109844])
* Exit gracefully when deleted {es-pull}107917[#107917] (issue: {es-issue}107266[#107266])
* Fix NPE during destination index creation {es-pull}108891[#108891] (issue: {es-issue}108890[#108890])
* Forward `indexServiceSafe` exception to listener {es-pull}108517[#108517] (issue: {es-issue}108418[#108418])
* Halt Indexer on Stop/Abort API {es-pull}107792[#107792]
* Handle `IndexNotFoundException` {es-pull}108394[#108394] (issue: {es-issue}107263[#107263])
* Prevent concurrent jobs during cleanup {es-pull}109047[#109047]
* Redirect `VersionConflict` to reset code {es-pull}108070[#108070]
* Reset max page size to settings value {es-pull}109449[#109449] (issue: {es-issue}109308[#109308])

Vector Search::
* Ensure vector similarity correctly limits `inner_hits` returned for nested kNN {es-pull}111363[#111363] (issue: {es-issue}111093[#111093])
* Ensure we return non-negative scores when scoring scalar dot-products {es-pull}108522[#108522]

Watcher::
* Avoiding running watch jobs in TickerScheduleTriggerEngine if it is paused {es-pull}110061[#110061] (issue: {es-issue}105933[#105933])

[[deprecation-8.15.0]]
[float]
=== Deprecations

ILM+SLM::
* Deprecate using slm privileges to access ilm {es-pull}110540[#110540]

Infra/Settings::
* `ParseHeapRatioOrDeprecatedByteSizeValue` for `indices.breaker.total.limit` {es-pull}110236[#110236]

Machine Learning::
* Deprecate `text_expansion` and `weighted_tokens` queries {es-pull}109880[#109880]

[[enhancement-8.15.0]]
[float]
=== Enhancements

Aggregations::
* Aggs: Scripted metric allow list {es-pull}109444[#109444]
* Enable inter-segment concurrency for low cardinality numeric terms aggs {es-pull}108306[#108306]
* Increase size of big arrays only when there is an actual value in the aggregators {es-pull}107764[#107764]
* Increase size of big arrays only when there is an actual value in the aggregators (Analytics module) {es-pull}107813[#107813]
* Optimise `BinaryRangeAggregator` for single value fields {es-pull}108016[#108016]
* Optimise cardinality aggregations for single value fields {es-pull}107892[#107892]
* Optimise composite aggregations for single value fields {es-pull}107897[#107897]
* Optimise few metric aggregations for single value fields {es-pull}107832[#107832]
* Optimise histogram aggregations for single value fields {es-pull}107893[#107893]
* Optimise multiterms aggregation for single value fields {es-pull}107937[#107937]
* Optimise terms aggregations for single value fields {es-pull}107930[#107930]
* Speed up collecting zero document string terms {es-pull}110922[#110922]

Allocation::
* Log shard movements {es-pull}105829[#105829]
* Support effective watermark thresholds in node stats API {es-pull}107244[#107244] (issue: {es-issue}106676[#106676])

Application::
* Add Create or update query rule API call {es-pull}109042[#109042]
* Rename rule query and add support for multiple rulesets {es-pull}108831[#108831]
* Support multiple associated groups for TopN {es-pull}108409[#108409] (issue: {es-issue}108018[#108018])
* [Connector API] Change `UpdateConnectorFiltering` API to have better defaults {es-pull}108612[#108612]

Authentication::
* Expose API Key cache metrics {es-pull}109078[#109078]

Authorization::
* Cluster state role mapper file settings service {es-pull}107886[#107886]
* Cluster-state based Security role mapper {es-pull}107410[#107410]
* Introduce role description field {es-pull}107088[#107088]
* [Osquery] Extend `kibana_system` role with an access to new `osquery_manager` index {es-pull}108849[#108849]

Data streams::
* Add metrics@custom component template to metrics-*-* index template {es-pull}109540[#109540] (issue: {es-issue}109475[#109475])
* Apm-data: enable plugin by default {es-pull}108860[#108860]
* Apm-data: ignore malformed fields, and too many dynamic fields {es-pull}108444[#108444]
* Apm-data: improve default pipeline performance {es-pull}108396[#108396] (issue: {es-issue}108290[#108290])
* Apm-data: improve indexing resilience {es-pull}108227[#108227]
* Apm-data: increase priority above Fleet templates {es-pull}108885[#108885]
* Apm-data: increase version for templates {es-pull}108340[#108340]
* Apm-data: set codec: best_compression for logs-apm.* data streams {es-pull}108862[#108862]
* Remove `default_field: message` from metrics index templates {es-pull}110651[#110651]

Distributed::
* Add `wait_for_completion` parameter to delete snapshot request {es-pull}109462[#109462] (issue: {es-issue}101300[#101300])
* Improve mechanism for extracting the result of a `PlainActionFuture` {es-pull}110019[#110019] (issue: {es-issue}108125[#108125])

ES|QL::
* Add `BlockHash` for 3 `BytesRefs` {es-pull}108165[#108165]
* Allow `LuceneSourceOperator` to early terminate {es-pull}108820[#108820]
* Check if `CsvTests` required capabilities exist {es-pull}108684[#108684]
* ESQL: Add aggregates node level reduction {es-pull}107876[#107876]
* ESQL: Add more time span units {es-pull}108300[#108300]
* ESQL: Implement LOOKUP, an "inline" enrich {es-pull}107987[#107987] (issue: {es-issue}107306[#107306])
* ESQL: Renamed `TopList` to Top {es-pull}110347[#110347]
* ESQL: Union Types Support {es-pull}107545[#107545] (issue: {es-issue}100603[#100603])
* ESQL: add REPEAT string function {es-pull}109220[#109220]
* ES|QL Add primitive float support to the Compute Engine {es-pull}109746[#109746] (issue: {es-issue}109178[#109178])
* ES|QL Add primitive float variants of all aggregators to the compute engine {es-pull}109781[#109781]
* ES|QL: vectorize eval {es-pull}109332[#109332]
* Optimize ST_DISTANCE filtering with Lucene circle intersection query {es-pull}110102[#110102] (issue: {es-issue}109972[#109972])
* Optimize for single value in ordinals grouping {es-pull}108118[#108118]
* Rewrite away type converting functions that do not convert types {es-pull}108713[#108713] (issue: {es-issue}107716[#107716])
* ST_DISTANCE Function {es-pull}108764[#108764] (issue: {es-issue}108212[#108212])
* Support metrics counter types in ESQL {es-pull}107877[#107877]
* [ESQL] CBRT function {es-pull}108574[#108574]
* [ES|QL] Convert string to datetime when the other size of an arithmetic operator is `date_period` or `time_duration` {es-pull}108455[#108455]
* [ES|QL] Support Named and Positional Parameters in `EsqlQueryRequest` {es-pull}108421[#108421] (issue: {es-issue}107029[#107029])
* [ES|QL] `weighted_avg` {es-pull}109993[#109993]

Engine::
* Drop shards close timeout when stopping node. {es-pull}107978[#107978] (issue: {es-issue}107938[#107938])
* Update translog `writeLocation` for `flushListener` after commit {es-pull}109603[#109603]

Geo::
* Optimize `GeoBounds` and `GeoCentroid` aggregations for single value fields {es-pull}107663[#107663]

Health::
* Log details of non-green indicators in `HealthPeriodicLogger` {es-pull}108266[#108266]

Highlighting::
* Unified Highlighter to support matched_fields  {es-pull}107640[#107640] (issue: {es-issue}5172[#5172])

Infra/Core::
* Add allocation explain output for THROTTLING shards {es-pull}109563[#109563]
* Create custom parser for ISO-8601 datetimes {es-pull}106486[#106486] (issue: {es-issue}102063[#102063])
* Extend ISO8601 datetime parser to specify forbidden fields, allowing it to be used on more formats {es-pull}108606[#108606]
* add Elastic-internal stable bridge api for use by Logstash {es-pull}108171[#108171]

Infra/Metrics::
* Add auto-sharding APM metrics {es-pull}107593[#107593]
* Add request metric to `RestController` to track success/failure (by status code) {es-pull}109957[#109957]
* Allow RA metrics to be reported upon parsing completed or accumulated {es-pull}108726[#108726]
* Provide the `DocumentSizeReporter` with index mode {es-pull}108947[#108947]
* Return noop instance `DocSizeObserver` for updates with scripts {es-pull}108856[#108856]

Ingest Node::
* Add `continent_code` support to the geoip processor {es-pull}108780[#108780] (issue: {es-issue}85820[#85820])
* Add support for the 'Connection Type' database to the geoip processor {es-pull}108683[#108683]
* Add support for the 'Domain' database to the geoip processor {es-pull}108639[#108639]
* Add support for the 'ISP' database to the geoip processor {es-pull}108651[#108651]
* Adding `hits_time_in_millis` and `misses_time_in_millis` to enrich cache stats {es-pull}107579[#107579]
* Adding `user_type` support for the enterprise database for the geoip processor {es-pull}108687[#108687]
* Adding human readable times to geoip stats {es-pull}107647[#107647]
* Include doc size info in ingest stats {es-pull}107240[#107240] (issue: {es-issue}106386[#106386])
* Make ingest byte stat names more descriptive {es-pull}108786[#108786]
* Return ingest byte stats even when 0-valued {es-pull}108796[#108796]
* Test pipeline run after reroute {es-pull}108693[#108693]

Logs::
* Introduce a node setting controlling the activation of the `logs` index mode in logs@settings component template {es-pull}109025[#109025] (issue: {es-issue}108762[#108762])
* Support index sorting with nested fields {es-pull}110251[#110251] (issue: {es-issue}107349[#107349])

Machine Learning::
* Add Anthropic messages integration to Inference API {es-pull}109893[#109893]
* Add `sparse_vector` query {es-pull}108254[#108254]
* Add model download progress to the download task status {es-pull}107676[#107676]
* Add rate limiting support for the Inference API {es-pull}107706[#107706]
* Add the rerank task to the Elasticsearch internal inference service {es-pull}108452[#108452]
* Default the HF service to cosine similarity {es-pull}109967[#109967]
* GA the update trained model action {es-pull}108868[#108868]
* Handle the "JSON memory allocator bytes" field {es-pull}109653[#109653]
* Inference Processor: skip inference when all fields are missing {es-pull}108131[#108131]
* Log 'No statistics at.. ' message as a warning {ml-pull}2684[#2684]
* Optimise frequent item sets aggregation for single value fields {es-pull}108130[#108130]
* Sentence Chunker {es-pull}110334[#110334]
* [Inference API] Add Amazon Bedrock Support to Inference API {es-pull}110248[#110248]
* [Inference API] Add Mistral Embeddings Support to Inference API {es-pull}109194[#109194]
* [Inference API] Check for related pipelines on delete inference endpoint {es-pull}109123[#109123]

Mapping::
* Add ignored field values to synthetic source {es-pull}107567[#107567]
* Apply FLS to the contents of `IgnoredSourceFieldMapper` {es-pull}109931[#109931]
* Binary field enables doc values by default for index mode with synthe… {es-pull}107739[#107739] (issue: {es-issue}107554[#107554])
* Feature/annotated text store defaults {es-pull}107922[#107922] (issue: {es-issue}107734[#107734])
* Handle `ignore_above` in synthetic source for flattened fields {es-pull}110214[#110214]
* Opt in keyword field into fallback synthetic source if needed {es-pull}110016[#110016]
* Opt in number fields into fallback synthetic source when doc values a… {es-pull}110160[#110160]
* Reflect latest changes in synthetic source documentation {es-pull}109501[#109501]
* Store source for fields in objects with `dynamic` override {es-pull}108911[#108911]
* Store source for nested objects {es-pull}108818[#108818]
* Support synthetic source for `geo_point` when `ignore_malformed` is used {es-pull}109651[#109651]
* Support synthetic source for `scaled_float` and `unsigned_long` when `ignore_malformed` is used {es-pull}109506[#109506]
* Support synthetic source for date fields when `ignore_malformed` is used {es-pull}109410[#109410]
* Support synthetic source together with `ignore_malformed` in histogram fields {es-pull}109882[#109882]
* Track source for arrays of objects {es-pull}108417[#108417] (issue: {es-issue}90708[#90708])
* Track synthetic source for disabled objects {es-pull}108051[#108051]

Network::
* Detect long-running tasks on network threads {es-pull}109204[#109204]

Ranking::
* Enabling profiling for `RankBuilders` and adding tests for RRF {es-pull}109470[#109470]

Relevance::
* [Query Rules] Add API calls to get or delete individual query rules within a ruleset {es-pull}109554[#109554]
* [Query Rules] Require Enterprise License for Query Rules {es-pull}109634[#109634]

Search::
* Add AVX-512 optimised vector distance functions for int7 on x64 {es-pull}109084[#109084]
* Add `SparseVectorStats` {es-pull}108793[#108793]
* Add `_name` support for top level `knn` clauses {es-pull}107645[#107645] (issues: {es-issue}106254[#106254], {es-issue}107448[#107448])
* Add a SIMD (AVX2) optimised vector distance function for int7 on x64 {es-pull}108088[#108088]
* Add min/max range of the `event.ingested` field to cluster state for searchable snapshots {es-pull}106252[#106252]
* Add per-field KNN vector format to Index Segments API {es-pull}107216[#107216]
* Add support for hiragana_uppercase & katakana_uppercase token filters in kuromoji analysis plugin {es-pull}106553[#106553]
* Adding support for explain in rrf {es-pull}108682[#108682]
* Allow rescorer with field collapsing {es-pull}107779[#107779] (issue: {es-issue}27243[#27243])
* Cut over stored fields to ZSTD for compression {es-pull}103374[#103374]
* Limit the value in prefix query {es-pull}108537[#108537] (issue: {es-issue}108486[#108486])
* Make dense vector field type updatable {es-pull}106591[#106591]
* Multivalue Sparse Vector Support {es-pull}109007[#109007]

Security::
* Add bulk delete roles API {es-pull}110383[#110383]
* Remote cluster - API key security model - cluster privileges {es-pull}107493[#107493]

Snapshot/Restore::
* Denser in-memory representation of `ShardBlobsToDelete` {es-pull}109848[#109848]
* Log repo UUID at generation/registration time {es-pull}109672[#109672]
* Make repository analysis API available to non-operators {es-pull}110179[#110179] (issue: {es-issue}100318[#100318])
* Track `RequestedRangeNotSatisfiedException` separately in S3 Metrics {es-pull}109657[#109657]

Stats::
* DocsStats: Add human readable bytesize {es-pull}109720[#109720]

TSDB::
* Optimise `time_series` aggregation for single value fields {es-pull}107990[#107990]
* Support `ignore_above` on keyword dimensions {es-pull}110337[#110337]

Vector Search::
* Adding hamming distance function to painless for `dense_vector` fields {es-pull}109359[#109359]
* Support k parameter for knn query {es-pull}110233[#110233] (issue: {es-issue}108473[#108473])

[[feature-8.15.0]]
[float]
=== New features

Aggregations::
* Opt `scripted_metric` out of parallelization {es-pull}109597[#109597]

Application::
* [Connector API] Add claim sync job endpoint {es-pull}109480[#109480]

ES|QL::
* ESQL: Add `ip_prefix` function {es-pull}109070[#109070] (issue: {es-issue}99064[#99064])
* ESQL: Introduce a casting operator, `::` {es-pull}107409[#107409]
* ESQL: `top_list` aggregation {es-pull}109386[#109386] (issue: {es-issue}109213[#109213])
* ESQL: add Arrow dataframes output format {es-pull}109873[#109873]
* Reapply "ESQL: Expose "_ignored" metadata field" {es-pull}108871[#108871]

Infra/REST API::
* Add a capabilities API to check node and cluster capabilities {es-pull}106820[#106820]

Ingest Node::
* Directly download commercial ip geolocation databases from providers {es-pull}110844[#110844]
* Mark the Redact processor as Generally Available {es-pull}110395[#110395]

Logs::
* Introduce logs index mode as Tech Preview {es-pull}108896[#108896] (issue: {es-issue}108896[#108896])

Machine Learning::
* Add support for Azure AI Studio embeddings and completions to the inference service. {es-pull}108472[#108472]

Mapping::
* Add `semantic_text` field type and `semantic` query {es-pull}110338[#110338]
* Add generic fallback implementation for synthetic source {es-pull}108222[#108222]
* Add synthetic source support for `geo_shape` via fallback implementation {es-pull}108881[#108881]
* Add synthetic source support for binary fields {es-pull}107549[#107549]
* Enable fallback synthetic source by default {es-pull}109370[#109370] (issue: {es-issue}106460[#106460])
* Enable fallback synthetic source for `point` and `shape` {es-pull}109312[#109312]
* Enable fallback synthetic source for `token_count` {es-pull}109044[#109044]
* Implement synthetic source support for annotated text field {es-pull}107735[#107735]
* Implement synthetic source support for range fields {es-pull}107081[#107081]
* Support arrays in fallback synthetic source implementation {es-pull}108878[#108878]
* Support synthetic source for `aggregate_metric_double` when ignore_malf… {es-pull}108746[#108746]

Ranking::
* Add text similarity reranker retriever {es-pull}109813[#109813]

Relevance::
* Mark Query Rules as GA {es-pull}110004[#110004]

Search::
* Add new int4 quantization to dense_vector {es-pull}109317[#109317]
* Adding RankFeature search phase implementation {es-pull}108538[#108538]
* Adding aggregations support for the `_ignored` field {es-pull}101373[#101373] (issue: {es-issue}59946[#59946])
* Update Lucene version to 9.11 {es-pull}109219[#109219]

Security::
* Query Roles API {es-pull}108733[#108733]

Transform::
* Introduce _transform/_node_stats API {es-pull}107279[#107279]

Vector Search::
* Adds new `bit` `element_type` for `dense_vectors` {es-pull}110059[#110059]

[[upgrade-8.15.0]]
[float]
=== Upgrades

Infra/Plugins::
* Update ASM to 9.7 for plugin scanner {es-pull}108822[#108822] (issue: {es-issue}108776[#108776])

Ingest Node::
* Bump Tika dependency to 2.9.2 {es-pull}108144[#108144]

Network::
* Upgrade to Netty 4.1.109 {es-pull}108155[#108155]

Search::
* Upgrade to Lucene-9.11.1 {es-pull}110234[#110234]

Security::
* Upgrade bouncy castle (non-fips) to 1.78.1 {es-pull}108223[#108223]

Snapshot/Restore::
* Bump jackson version in modules:repository-azure {es-pull}109717[#109717]