This documentation contains work-in-progress information for future Elastic Stack and Cloud releases. Use the version selector to view supported release docs. It also contains some Elastic Cloud serverless information. Check out our serverless docs for more details.
Elasticsearch version 8.8.0
editElasticsearch version 8.8.0
editAlso see Breaking changes in 8.8.
Bug fixes
edit- Aggregations
- Allocation
-
- Avoid copying during iteration of all shards in routing table #94417
-
Avoid duplicate application of
RoutingTable
diff #94379 - Balance priorities during reconciliation #95454
-
Fix
RebalanceOnlyWhenActiveAllocationDecider
#96025 -
Streamline
AsyncShardFetch#getNumberOfInFlightFetches
#93632 (issue: #93631)
- Application
- Authorization
- CRUD
-
-
Avoid null
Location
in post write refresh #95229
-
Avoid null
- Cluster Coordination
-
-
Read register current term asynchronously in
StoreHeartbeatService
#95351
-
Read register current term asynchronously in
- DLM
- Data streams
- Distributed
- Engine
-
- Ensure refresh to return the latest commit generation #94249
- Geo
- ILM+SLM
- Infra/Core
- Infra/REST API
- Infra/Scripting
- Ingest Node
- Machine Learning
-
- Avoid expensive source parsing by using doc values when querying model definition meta fields #95590
- Mapping
-
- Longer timeout for mapping update during resize #95221
- Network
- Ranking
-
- Fix versioning for tests cases using a randomly generated rank builder #95514
- Search
- Snapshot/Restore
- Transform
-
- Call listener in order to prevent the request from hanging #96221
-
Do not fail upon
ResourceAlreadyExistsException
during destination index creation #96274 (issue: #95310) -
Fix privileges check failures by adding
allow_restricted_indices
flag #95187 - Secondary credentials used with transforms should only require source and destination index privileges, not transform privileges #94420
-
Use monotonic time in
TransformScheduler
#95456 (issue: #95445)
Deprecations
edit- Allocation
-
-
Deprecate
cluster.routing.allocation.type
#94066
-
Deprecate
Enhancements
edit- Aggregations
- Analysis
-
- Add origin of synonym rules to exception message #93702
- Application
- Authentication
- Authorization
- Cluster Coordination
- DLM
-
- Adding origination date to DLM #95113
- Engine
-
- Increase the merge factor to 32 for time-based data #94134
- Reduce the likelihood of writing small segments due to an oversize translog #93524 (issue: #75611)
- Sort segments on timestamp in read only engine #93576
-
Use
LogByteSizeMergePolicy
instead ofTieredMergePolicy
for time-based data #92684 - Use mmap for temporary files #93595
- Geo
- Highlighting
-
-
Use
storedFieldsSpec
to load stored fields for highlighting #91841
-
Use
- ILM+SLM
- Indices APIs
- Infra/Core
-
- Allow preserving specific headers on thread context stash #94680
- Infra/Plugins
- Infra/Transport API
-
-
Add
transport_version
to node info JSON #94669
-
Add
- Ingest Node
- Machine Learning
-
-
Add
_meta
field to data frame analytics config #94529 -
Add
embedding_size
to text embedding config #95176 - Include model definition install status for Pytorch models #95271
- Integrate ELSER model download into put trained model API #95281
- Start, stop and infer of a trained model can now optionally use a deployment ID that is different to the model ID #95168
- [ML] Get trained model stats by deployment id or model id #95440
-
Add
- Mapping
- Monitoring
-
-
Add
event_loop_utilization
Kibana stats to the monitoring index templates #95388
-
Add
- Network
- SQL
-
-
Add
WildcardLike/Pattern
to QL #95357
-
Add
- Search
-
- Adding initial public and internal serverless scopes to Search team REST handlers #94035
-
Enable
_terms_enum
on version fields #93839 (issue: #83403) -
Introduce
DocumentParsingException
#92646 (issue: #85083) -
Leverage
Weight#count
when size is set to 0 #94858 -
Make
SourceProvider
using stored fields segment-thread-safe #95082 -
Shortcut total hit count when
terminate_after
is used #94889 - [Profiling] Map stack frames more efficiently #94327
- [Profiling] Parallelize response handling #93960
- Security
-
-
Fleet: Add new mappings for
.fleet-actions
signing #93802
-
Fleet: Add new mappings for
- Snapshot/Restore
- TSDB
-
-
Support position
time_series_metric
ongeo_point
fields #93946
-
Support position
- Transform
- Vector Search
-
- Increase max number of vector dims to 2048 #95257
- Watcher
New features
edit- Application
- Authentication
-
- GA release of the JWT realm #95398
- CRUD
-
-
New
TransportBroadcastUnpromotableAction
action #93600
-
New
- DLM
-
- Add new endpoints to configure data lifecycle on a data stream level #94590
- Dlm add auto rollover condition max age #94950
-
Initial implementation for
DataLifecycleService
#94012 - Introduce a _lifecycle/explain API for data stream backing indices #94621
-
Introduce the
index.lifecycle.prefer_ilm
setting #95423 - [DLM] Extend the template to simulate api to support include defaults #94861
- [DLM] Introduce default rollover cluster setting & expose it via APIs #94240
- Health
- Ingest Node
-
- Add license checking to the redact processor #95477
- Machine Learning
-
- Text Expansion Query #93694
- Ranking
-
- Add support for Reciprocal Rank Fusion to the search API #93396
- Search
- TSDB