8.17.1

8.17.1 (2025-01-23)

• Make pyarrow dependency optional for tests (#2733, contributed by @danigm)
• Update APIs:
  • Add Simulate ingest API
  • Add Get data stream lifecycle stats API
  • Add Update inference API
  • Add Create or update, Get and Delete IP geolocation database configuration APIs
  • Add Bulk update API keys
  • Add Get and Update Security index settings APIs
  • Add OpenID Connect prepare authentication, OpenID Connect authenticate and OpenID Connect logout APIs
  • Add Delegate PKI authentication API
  • Add Repository analysis API
  • Add Render Search Application Query API
  • Add Find field structure and Find messages structure APIs
  • Add Get Watcher index settings and Update Watcher index settings APIs
  • Add experimental Check in and Claim connector sync job APIs
  • Add experimental Set connector sync job errors and Set connector sync job stats APIs
  • Add experimental Update connector features APIs
  • Add experimental Post Event to an Analytics Collection API
  • Add timeout and master_timeout to Snapshot lifecycle management (SLM) APIs
  • Add allow_partial_search_results to SQL search API
  • Add throttle_period_in_millis to Create or update watch API
  • Fix query parameters for CAT APIs

8.17.0

8.17.0 (2024-12-13)

• Allow simsimd again on Python 3.13 (#2722)
• Update APIs:
  • Mark all Inference APIs as stable.
  • Add allow_partial_search_results to the Open Point in Time API
  • Add keep_alive to the Get async search status API
  • Remove the keep_alive, pre_filter_shard_size and scroll parameters from the Submit async search API. They were never supported.
  • Add master_timeout and timeout to all autoscaling policy APIs
  • Add master_timeout to the Alias exists and Get alias APIs
  • Add list_executed_pipelines and require_data_stream to Bulk API
  • Add include_model_definition to Get trained models API
  • Add meta to Create data frame analytics API
  • Add aggs to Create datafeeds API
  • Add allow_no_indices, expand_wildcards, ignore_throttled and ignore_unavailable to Create anomaly detection jobs API

8.16.0

8.16.0 (2024-11-12)

• Support Python 3.13 (#2680)
• Emit Python warnings for beta and tech preview APIs (#2675). You can disable those warnings if needed:

  import warnings
  from elasticsearch.exceptions import GeneralAvailabilityWarning
  
  warnings.filterwarnings("ignore", category=GeneralAvailabilityWarning)

• Vectorstore: use a retriever query for hybrid search (#2666)
• Allow retries for statuses other than 429 in streaming bulk (#2071, contributed by @ayayron)
• Make BulkIndexError and ScanError serializable (#2669, contributed by @seagrine)
• Fix import when trace is missing from opentelemetry package (#2694, contributed by @nicoloboschi)
• Update APIs:
  • Fix nodes parameter in Task management API (contributed by @margaretpearce)
  • Add Test query rule API
  • Add Create Cross-Cluster API key and Update Cross-Cluster API key APIs
  • Add experimental Verify snapshot repository API
  • Add data_stream_name and settings to Delete auto-follow pattern API
  • Add max_samples_per_key to Get async EQL status API
  • Add lifecycle and remove unused data_retention and downsampling parameters from Put data stream lifecycle API
  • Add include_remotes and remove flat_settings from Cluster stats API
  • Add remote_indices to Create or update application privileges and Create or update roles APIs

7.17.12

• Fix compatibility with NumPy 2.0 (#2647, contributed by @alvarogf97)

8.15.1

8.15.1 (2024-09-08)

• Fix OTel context loss in parallel bulk helper (#2616)
• Use request converter to generate python examples (#2645)
• Add Geoip database configuration APIs: Create or update, Get and Delete
• Add q parameter to Update by Query API
• Add allow_no_indices and ignore_unavailable parameters to Resolve index API

8.15.0

• Added the Connector API (#2623)
• Added support for semantic_text and semantic query.
• Added support for sequences of job id, model id and tags where applicable in ML APIs
• Added dry_run and force parameters to the Perform inference API
• Added optional Arrow deserialization support (#2632)
• Merged Query Ruleset API into new Query Rules API (#2607)
• Added mapping code examples (#2596)
• Fixed reference docs (#2629)
• Dropped Python 3.7 support (#2618)

8.14.0

• Fixed node_pool_class override (#2581, contributed by @tallakh)
• Added retriever to the Search API
• Added deprecated and removed allow_auto_create from the Create or update component template API
• Added allow_auto_create, cause, deprecated, ignore_missing_component_templates and master_timeout to the Create or update index template API
• Added cause, removed flat_settings and timeout from the Create or update index template legacy API
• Removed various unsupported parameters from the Simulate index API
• Added various supported paramters to the Simulate index template API
• Added the completion and rerank task types to the Inference APIs
• Added the query and timeout parameters to the Perform inference API
• Added typed_keys to the Search Application Search API
• Added with_profile_uid to the Get API key information and Query API key information APIs

8.13.2

• Added the ml.update_trained_model_deployment API
• Marked Requests 2.32.2 as incompatible with the Elasticsearch client

8.13.1

• Added force_synthetic_source to the Get API
• Added wait_for_completion to the Create trained model API
• Added typed_keys to the Query API key information API

8.13.0

• Added native OpenTelemetry support (#2435)
• Added optional orjson (a fast, correct JSON library) serialization support (#2493)
• Added the text_structure.test_grok_pattern API
• Added the indices.resolve_cluster API
• Renamed the model_id parameter to inference_id in the inference APIs
• Changed all synonyms APIs from experimental to stable.
• Fixed API key documentation (#2477, contributed by @iuliaferoli)