NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #553 most downloaded on PyPI
Python client for Elasticsearch
Last release 21 days ago
09 Sep 2026
Ships fairly regularly
a new release about every 6 weeks
Nearly every release is documented
notes for 59 of the last 60 stable releases
2 versions withdrawn
withdrawn after publishing
13 years old
174 releases · first in 2013
Client is compatible with Elasticsearch 8.2.2
Added the inference_config parameter to the ml.infer_trained_model_deployment API.
inference_config parameter to the ml.infer_trained_model_deployment API.Re-introduced support for passing requests.auth.BaseAuth objects to the http_auth parameter which was available in 7.x.
One column per quarter.
requests.auth.BaseAuth objects to the http_auth parameter which was available in 7.x.filter parameter to the experimental knn_search APIsource and dest parameters for the reindex API from optional to requiredindices.field_usage_stats APIindices.modify_data_stream APIfields and types parameters to the field_caps APIignore_unvailable parameter to the open_point_in_time APImaster_timeout and timeout parameters to the indices.delete APIfeatures parameter to the indices.get APIml.get_memory_stats APImigrations.get_feature_upgrade_status APImigrations.post_feature_upgrade APInodes.clear_repositories_metering_archive APInodes.get_repositories_metering_info APIsecurity.activate_user_profile APIsecurity.disable_user_profile APIsecurity.enable_user_profile APIsecurity.get_user_profile APIsecurity.suggest_user_profiles APIsecurity.update_user_profile_data APIcatalog, index_using_frozen, keep_alive, keep_on_completion, runtime_mappings, and wait_for_completion_timeout parameters to the sql.query APIClient is compatible with Elasticsearch 8.1.3
Client is compatible with Elasticsearch 8.1.2
Changed the source and dest parameters of the reindex API to be required.
source and dest parameters of the reindex API to be required.fields parameter of the field_caps API to be required.Added the transform.reset_transform API
transform.reset_transform APIFixed the sort parameter of multiple APIs to serialize in the query string when using the shorthand form
sort parameter of multiple APIs to serialize in the query string when using the shorthand formindex parameter to be required for the indices.field_usage_stats APIcatalog, keep_alive, keep_on_completion, params, runtime_mappings, and wait_for_completion_timeout parameters to the sql.query API:warning: v8.0.0 is a new major release and contains some breaking changes. For information on how migrate your application from v7.x to v8.0.0 you ca…
:warning: v8.0.0 is a new major release and contains some breaking changes. For information on how migrate your application from v7.x to v8.0.0 you can read the migration guide here.
.options() method to Elasticsearch and AsyncElasticsearch for modifying transport options.basic_auth parameter for specifying username and password authenticationbearer_auth parameter for specifying an HTTP bearer token or service tokenmeta property to ApiError to access the HTTP response metadata of an error.elastic-transport package is installed.elastic-transport packagebody parameters to have semantic names (e.g index(document={...}) instead of index(body={...})).meta for response metadata (HTTP status, headers, node, etc) and body for a typed body.AsyncElasticsearch to always be available, regardless of whether aiohttp is installedApiError which differentiates between an error that's raised from the transport layer (previously elasticsearch.exceptions.TransportError, now elastic_transport.TransportError) and one raised from the API layerJSONSerializer to JsonSerializer for consistency with other serializer names. Added an alias to the old name for backwards compatibilityapplication/json) to instead use compatibility mimetypes (application/vnd.elasticsearch+json) which always request for responses compatibility with version 8.x.elasticsearch.connection module as all functionality has been moved to the elastic-transport packagehttp://localhost:9200 due to Elasticsearch 8.0 default configuration being https://localhost:9200.
The client's connection to Elasticsearch now must be specified with scheme, host, and port or with the cloud_id parameterbody and params parameters on all APIshttp_auth, api_key, ignore, request_timeout, headers, and opaque_id
All of these settings should instead be set via the .options() methodelasticsearch.transport and elasticsearch.client modules. These modules will be removed in a future versiondoc_type, include_type_name, and copy_settings parameters from many document and index APIslocal parameter from the cat.indices, cat.nodes, cat.shards APIallow_no_datafeeds parameter from the cat.ml_datafeeds APIallow_no_jobs parameter from the cat.ml_jobs APIsize parameter from the cat.thread_pool APItime parameter to the cat.thread_pool APIsize parameter from the delete_by_query APIsize parameter from the update_by_query APIindices.flush_synced APIindices.freeze APIindices.get_upgrade APIindices.upgrade APIindices.exist_type APIcopy_settings from the indices.shrink APIverbose parameter of the indices.segments APIaccept_enterprise parameter of the license.get APIaccept_enterprise parameter of the xpack.info APIml.infer_trained_model_deployment APIml.put_trained_model_definition_part APIml.put_trained_model_vocabulary APIml.start_trained_model_deployment APIml.stop_trained_model_deployment APItimeout parameter to the ml.delete_trained_model APIallow_no_jobs parameter from the ml.close_job APIml.find_text_structure APIallow_no_datafeeds parameter from the ml.get_datafeed_stats APIallow_no_datafeeds parameter from the ml.get_datafeeds APIallow_no_jobs parameter from the ml.get_job_stats APIallow_no_jobs parameter from the ml.get_jobs APIallow_no_jobs parameter from the ml.get_overall_buckets APIknn_search APIsearchable_snapshots.repository_stats APIsnapshot.delete API to accept multiple snapshotssecurity.enroll_kibana APIsecurity.enroll_node APIDeprecated the body and params parameters on all APIs
.options() method to Elasticsearch and AsyncElasticsearch for modifying transport options.basic_auth parameter for specifying username and password authenticationbearer_auth parameter for specifying an HTTP bearer token or service tokenmeta property to ApiError to access the HTTP response metadata of an error.elastic-transport package is installed.elastic-transport packagebody parameters to have semantic names (e.g index(document={...}) instead of index(body={...})).meta for response metadata (HTTP status, headers, node, etc) and body for a typed body.AsyncElasticsearch to always be available, regardless of whether aiohttp is installedApiError which differentiates between an error that's raised from the transport layer (previously elasticsearch.exceptions.TransportError, now elastic_transport.TransportError) and one raised from the API layerJSONSerializer to JsonSerializer for consistency with other serializer names. Added an alias to the old name for backwards compatibilityapplication/json) to instead use compatibility mimetypes (application/vnd.elasticsearch+json) which always request for responses compatibility with version 8.x.elasticsearch.connection module as all functionality has been moved to the elastic-transport packagehttp://localhost:9200 due to Elasticsearch 8.0 default configuration being https://localhost:9200.
The client's connection to Elasticsearch now must be specified with scheme, host, and port or with the cloud_id parameterbody and params parameters on all APIshttp_auth, api_key, ignore, request_timeout, headers, and opaque_id
All of these settings should instead be set via the .options() methodelasticsearch.transport and elasticsearch.client modules. These modules will be removed in a future versiondoc_type and include_type_name parameters from many document and index APIslocal parameter from the cat.indices, cat.nodes, cat.shards APIallow_no_datafeeds parameter from the cat.ml_datafeeds APIallow_no_jobs parameter from the cat.ml_jobs APIsize parameter from the cat.thread_pool APItime parameter to the cat.thread_pool APIsize parameter from the delete_by_query APIsize parameter from the update_by_query APIindices.flush_synced APIindices.freeze APIindices.get_upgrade APIindices.upgrade APIcopy_settings from the indices.shrink APIaccept_enterprise parameter of the license.get APIaccept_enterprise parameter of the xpack.info APIallow_no_jobs parameter from the ml.close_job APItimeout parameter to the ml.delete_trained_model APIml.find_text_structure APIallow_no_datafeeds parameter from the ml.get_datafeed_stats APIallow_no_datafeeds parameter from the ml.get_datafeeds APIallow_no_jobs parameter from the ml.get_job_stats APIallow_no_jobs parameter from the ml.get_jobs APIallow_no_jobs parameter from the ml.get_overall_buckets APIml.infer_trained_model_deployment APIml.put_trained_model_definition_part APIml.put_trained_model_vocabulary APIml.start_trained_model_deployment APIml.stop_trained_model_deployment APIknn_search APIsearchable_snapshots.repository_stats APIsecurity.enroll_kibana APIsecurity.enroll_node APIDeprecated the body and params parameters on all APIs
.options() method to Elasticsearch and AsyncElasticsearch for modifying transport options.basic_auth parameter for specifying username and password authenticationbearer_auth parameter for specifying an HTTP bearer token or service tokenmeta property to ApiError to access the HTTP response metadata of an error.elastic-transport package is installed.elastic-transport packagebody parameters to have semantic names (e.g index(document={...}) instead of index(body={...})).meta for response metadata (HTTP status, headers, node, etc) and body for a typed body.AsyncElasticsearch to always be available, regardless of whether aiohttp is installedApiError which differentiates between an error that's raised from the transport layer (previously elasticsearch.exceptions.TransportError, now elastic_transport.TransportError) and one raised from the API layerJSONSerializer to JsonSerializer for consistency with other serializer names. Added an alias to the old name for backwards compatibilityelasticsearch.connection module as all functionality has been moved to the elastic-transport packagehttp://localhost:9200 due to Elasticsearch 8.0 default configuration being https://localhost:9200.
The client's connection to Elasticsearch now must be specified with scheme, host, and port or with the cloud_id parameterbody and params parameters on all APIshttp_auth, api_key, ignore, request_timeout, headers, and opaque_id
All of these settings should instead be set via the .options() methodelasticsearch.transport and elasticsearch.client modules. These modules will be removed in a future versiondoc_type and include_type_name parameters from many document and index APIslocal parameter from the cat.indices, cat.nodes, cat.shards APIallow_no_datafeeds parameter from the cat.ml_datafeeds APIallow_no_jobs parameter from the cat.ml_jobs APIsize parameter from the cat.thread_pool APItime parameter to the cat.thread_pool APIsize parameter from the delete_by_query APIsize parameter from the update_by_query APIindices.flush_synced APIindices.freeze APIindices.get_upgrade APIindices.upgrade APIcopy_settings from the indices.shrink APIaccept_enterprise parameter of the license.get APIaccept_enterprise parameter of the xpack.info APIallow_no_jobs parameter from the ml.close_job APItimeout parameter to the ml.delete_trained_model APIml.find_text_structure APIallow_no_datafeeds parameter from the ml.get_datafeed_stats APIallow_no_datafeeds parameter from the ml.get_datafeeds APIallow_no_jobs parameter from the ml.get_job_stats APIallow_no_jobs parameter from the ml.get_jobs APIallow_no_jobs parameter from the ml.get_overall_buckets APIml.infer_trained_model_deployment APIml.put_trained_model_definition_part APIml.put_trained_model_vocabulary APIml.start_trained_model_deployment APIml.stop_trained_model_deployment APIknn_search APIsearchable_snapshots.repository_stats APIsecurity.enroll_kibana APIsecurity.enroll_node APIDeprecated setting transport options http_auth, api_key, ignore, request_timeout, headers, and opaque_id All of these settings should instead be set v…
.options() method to Elasticsearch and AsyncElasticsearch for modifying transport options.basic_auth parameter for specifying username and password authenticationbearer_auth parameter for specifying an HTTP bearer token or service tokenmeta property to ApiError to access the HTTP response metadata of an error.elastic-transport package is installed.elastic-transport packagebody parameters to have semantic names (e.g index(document={...}) instead of index(body={...} [elasticsearch8-8.0.0a3.tar.gz](https://github.com/elastic/elasticsearch-py/files/7686727/elasticsearch8-8.0.0a3.tar.gz) ).meta for response metadata, raw for the raw deserialized response, and body for a typed body.AsyncElasticsearch to always be available, regardless of whether aiohttp is installedApiError which differentiates between an error that's raised from the transport layer (previously elasticsearch.exceptions.TransportError, now elastic_transport.TransportError) and one raised from the API layerJSONSerializer to JsonSerializer for consistency with other serializer names. Added an alias to the old name for backwards compatibilityelasticsearch.connection module as all functionality has been moved to the elastic-transport packagehttp://localhost:9200 due to Elasticsearch 8.0 default configuration being https://localhost:9200.
The client's connection to Elasticsearch now must be specified with scheme, host, and port or with the cloud_id parameterhttp_auth, api_key, ignore, request_timeout, headers, and opaque_id
All of these settings should instead be set via the .options() methodelasticsearch.transport and elasticsearch.client modules. These modules will be removed in a future versionbody and params parameters on all APIsdoc_type and include_type_name parameters from many document and index APIslocal parameter from the cat.indices, cat.nodes, cat.shards APIallow_no_datafeeds parameter from the cat.ml_datafeeds APIallow_no_jobs parameter from the cat.ml_jobs APIsize parameter from the cat.thread_pool APItime parameter to the cat.thread_pool APIsize parameter from the delete_by_query APIsize parameter from the update_by_query APIindices.flush_synced APIindices.freeze APIindices.get_upgrade APIindices.upgrade APIcopy_settings from the indices.shrink APIaccept_enterprise parameter of the license.get APIaccept_enterprise parameter of the xpack.info APIallow_no_jobs parameter from the ml.close_job APItimeout parameter to the ml.delete_trained_model APIml.find_text_structure APIallow_no_datafeeds parameter from the ml.get_datafeed_stats APIallow_no_datafeeds parameter from the ml.get_datafeeds APIallow_no_jobs parameter from the ml.get_job_stats APIallow_no_jobs parameter from the ml.get_jobs APIallow_no_jobs parameter from the ml.get_overall_buckets APIml.infer_trained_model_deployment APIml.put_trained_model_definition_part APIml.put_trained_model_vocabulary APIml.start_trained_model_deployment APIml.stop_trained_model_deployment APIknn_search APIsearchable_snapshots.repository_stats APIsecurity.enroll_kibana APIsecurity.enroll_node APINothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Fix compatibility with NumPy 2.0 (#2647, contributed by @alvarogf97)
Client is compatible with Elasticsearch 7.17.9
Removed the work-around for a bug in aiohttp when sending HEAD requests. This change was backported from v8.x of elastic-transport.
Client is compatible with Elasticsearch 7.17.7
Client is compatible with Elasticsearch 7.17.6
Client is compatible with Elasticsearch 7.17.5
Client is compatible with Elasticsearch 7.17.4
Client is compatible with Elasticsearch 7.17.3
Client is compatible with Elasticsearch 7.17.2
Fixed the from parameter to be rewritten to from_ when used with the scan and async_scan helpers
from parameter to be rewritten to from_ when used with the scan and async_scan helpers (https://github.com/elastic/elasticsearch-py/pull/1897)AiohttpHttpConnection to not leak TLS connections when the socket isn't explicitly shutdown by the peer (https://github.com/elastic/elasticsearch-py/pull/1919)Added the ml.get_model_snapshot_upgrade_stats API
ml.get_model_snapshot_upgrade_stats APIbody parameter to the ml.forecast and ml.open_job APIstimeout parameter to the transform.delete_transform, transform.preview_transform, transform.put_transform, transform.update_transform, and transform.upgrade_transform APIsAPI is compatible with Elasticsearch 7.16.3
API is compatible with Elasticsearch 7.16.2
API is compatible with Elasticsearch 7.16.1
AIOHttpConnection wouldn't log query parameters for URLs.Deprecated the send_get_body_as parameter. This parameter is no longer necessary as APIs all use non-GET HTTP methods when using a body.
send_get_body_as parameter. This parameter is no longer necessary
as APIs all use non-GET HTTP methods when using a body.body, params, and other per-request parameters has been delayed beyond 8.0.0.
Changed deprecation warnings to mention "future version" instead of 8.0.0.unicode HTTP headers with the urllib3 HTTP clientscan helper to always set the sort and scroll parameterskeep_alive parameter of the open_point_in_time API to be required
to reflect its required status within Elasticsearchtrack_total_hits parameter to the search_mvt APIfleet.global_checkpoints API from experimental to stablefleet.search experimental APIfleet.msearch experimental APIindices.modify_data_stream APIif_version parameter to the ingest.put_pipeline APImigration.get_feature_upgrade_status APImigration.post_feature_upgrade APIdefer_definition_decompression parameter to the ml.put_trained_model APItransform.upgrade_transforms APIAPI is compatible with Elasticsearch 7.15.2
metric parameter of the nodes.info API.Changed the DeprecationWarning for the body parameter to be a "removed in a future version" instead of "removed in 8.0" in line with the 8.0 roadmap.
JSONSerializer.default() when numpy and pandas weren't installed.DeprecationWarning for the body parameter to be a "removed in a future version" instead of "removed in 8.0" in line with the 8.0 roadmap.index parameter of the open_point_in_time API is now required, was optional.We've recently announced our roadmap for 8.0.0 and have begun issuing deprecation warnings in the 7.x release stream to prepare users for upgrading to…
We've recently announced our roadmap for 8.0.0 and have begun issuing deprecation warnings in the 7.x release stream to prepare users for upgrading to 8.0.0 when available.
AsyncTransport and AIOHttpConnectionMapboxVectorTileSerializer for handling the application/vnd.mapbox-vector-tile mimetype. Because this mimetype is binary rather than text the raw response bytes are forwarded from the serializer without decodingelasticsearch module by delaying imports of pandas and numpy until later in the JSON serialization stage if necessarysearch_mvt experimental APIsearch, scroll, and clear_scroll APIsbody parameter of the search, scroll, and clear_scroll APIsupdate APIdocument parameter to the create and index APIsbody parameter of the create, index, and update APIsindices.disk_usage experimental APIindices.fields_usage_stats experimental APIindices.create APIbody parameter of the indices.create APIignore_unavailable, allow_no_indices, ignore_throttled, and expand_wildcards parameters to the ml.put_job APInodes.clear_repositories_metering_archive experimental APInodes.get_repositories_metering_info experimental APIshards option to the index_metric parameter of the nodes.stats APIdoc_type parameter of the nodes.hot_threads API, instead use the type parametersecurity.query_api_keys APIdoc_type parameter of the license.post_start_trial API, instead use the type parameterWe've recently announced our roadmap for 8.0.0 and have begun issuing deprecation warnings in the 7.x release stream to prepare users for upgrading to…
We've recently announced our roadmap for 8.0.0 and have begun issuing deprecation warnings in the 7.x release stream to prepare users for upgrading to 8.0.0 when available.
AsyncTransport and AIOHttpConnectionMapboxVectorTileSerializer for handling the application/vnd.mapbox-vector-tile mimetype. Because this mimetype is binary rather than text the raw response bytes are forwarded from the serializer without decoding.search_mvt experimental APIsearch, scroll, and clear_scroll APIsbody parameter of the search, scroll, and clear_scroll APIsupdate APIdocument parameter to the create and index APIsbody parameter of the create, index, and update APIsindices.disk_usage experimental APIindices.fields_usage_stats experimental APIindices.create APIbody parameter of the indices.create APIignore_unavailable, allow_no_indices, ignore_throttled, and expand_wildcards parameters to the ml.put_job API.nodes.clear_repositories_metering_archive experimental APInodes.get_repositories_metering_info experimental APIshards option to the index_metric parameter of the nodes.stats APIdoc_type parameter of the nodes.hot_threads API, instead use the type parametersecurity.query_api_keys APIdoc_type parameter of the license.post_start_trial API, instead use the type parameterClient is compatible with Elasticsearch 7.14.2
Client is compatible with Elasticsearch 7.14.1
Added check that client is connected to an Elasticsearch cluster. If the client isn't connected to a supported Elasticsearch cluster the UnsupportedPr
UnsupportedProductError exception will be raised.Added the terms_enum beta API
Removed the query_and_fetch and dfs_query_and_fetch options in the search_type parameter to the msearch, msearch_template and search_template APIs
ilm.migrate_to_data_tiers APIml.reset_job APIsecurity.saml_authenticate APIsecurity.saml_complete_logout APIsecurity.saml_invalidate APIsecurity.saml_logout APIsecurity.saml_prepare_authentication APIsecurity.saml_service_provider_metadata APIsql.delete_async APIsql.get_async APIsql.get_async_status APIinclude_repository parameter to snapshot.get APIrarely_abort_writes parameter to the snapshot.repository_analyze APINothing published for this version
Added check that client is connected to an Elasticsearch cluster. If the client isn't connected to Elasticsearch a NotElasticsearchError exception wil
NotElasticsearchError exception will be raised.Added the terms_enum beta API
Removed the query_and_fetch and dfs_query_and_fetch options in the search_type parameter to the msearch, msearch_template and search_template APIs
ilm.migrate_to_data_tiers APIml.reset_job APIsecurity.saml_authenticate APIsecurity.saml_complete_logout APIsecurity.saml_invalidate APIsecurity.saml_logout APIsecurity.saml_prepare_authentication APIsecurity.saml_service_provider_metadata APIinclude_repository parameter to snapshot.get APIrarely_abort_writes parameter to the snapshot.repository_analyze APIClient is compatible with Elasticsearch 7.13.4
NameError would be raised on Python 2.7 and 3.4 when a connection error would have otherwise been raised.
NameError would be raised on Python 2.7 and 3.4 when a connection error would have otherwise been raised.:warning: This release has been yanked on PyPI due to a regression in Python <3.5, please use another version.
:warning: This release has been yanked on PyPI due to a regression in Python <3.5, please use another version.
Transport.perform_request() now properly reraises RecursionError (Contributed by @hmilkovi)AIOHttpConnection no longer sends Accept-Encoding: gzip, deflate when http_compress=None.
Instead now sends no Accept-Encoding header.snapshot.repository_analyze APIClient is compatible with Elasticsearch 7.13.1
Added support for compatibility header for Elasticsearch. If the environment variable ELASTIC_CLIENT_APIVERSIONING=1 is set the client will send the h
ELASTIC_CLIENT_APIVERSIONING=1 is set the client will send the headers Accept and Content-Type with the following value: application/vnd.elasticsearch+json;compatible-with=7.include_unloaded_segments parameter to the cat.nodes APIfeatures.reset_features experimental APIfleet.global_checkpoints expiremental APIingest.geo_ip_stats APIml.delete_trained_model_alias APIml.preview_data_frame_analytics APIml.put_trained_model_alias APIml.delete_data_frame_analytics, ml.delete_trained_model, ml.explain_data_frame_analytics, ml.get_data_fram_analytics, ml.get_data_frame_analytics_stats, ml.get_trained_models, ml.get_trained_models_stats, ml.put_trained_model, ml.start_data_frame_analytics, ml.stop_data_frame_analytics, ml.update_data_frame_analytics APIs from beta to stable.include_unloaded_segments parameter to node.stats APIsearchable_snapshots.cache_stats experimental APIsecurity.clear_cached_service_tokens beta APIsecurity.create_service_token beta APIsecurity.delete_service_token beta APIsecurity.get_service_accounts beta APIsecurity.get_service_credentials beta APIshutdown.delete_node experiemental APIshutdown.get_node experimental APIshutdown.put_node experimental APIindex_details parameter to snapshot.get APItext_structure.find_structure API from experimental to stableChanged the text_structure.find_text_structure API from experimental to stable
text_structure.find_text_structure API from experimental to stableChanged autoscaling.delete_autoscaling_policy, autoscaling.get_autoscaling_policy, and autoscaling.put_autoscaling_policy APIs from experimental to st
autoscaling.delete_autoscaling_policy, autoscaling.get_autoscaling_policy, and autoscaling.put_autoscaling_policy APIs from experimental to stableeql.get_status APIlogstash.delete_pipeline, logstash.get_pipeline, and logstash.put_pipeline APIsml.find_text_structure APIstorage parameter to the searchable_snapshots.mount APIlevel parameter to the searchable_snapshots.stats APImin_compatible_shard_node parameter to search()text_structure.find_text_structure APIAdded ElasticsearchWarning which is raised when the Warning HTTP header is returned from Elasticsearch. ElasticsearchDeprecationWarning is now an alia…
X-Elastic-Client-Meta HTTP header and the meta_header parameter for controlling the header (#1473)ElasticsearchWarning which is raised when the Warning HTTP header is returned from Elasticsearch. ElasticsearchDeprecationWarning is now an alias for this warning type (#1495)async_search.status APIautoscaling.get_autoscaling_capacity experimental APIautoscaling.get_autoscaling_decision experimental APIcat.tasks API parameters node_id and parent_task to nodes and parent_task_idcluster.delete_component_template, cluster.exists_component_template, and cluster.get_component_template,
cluster.put_component_template APIs from experimental to stableeql.delete, eql.get, eql.search APIs from beta to stableindices.migrate_to_data_stream APIindices.promote_data_stream APIexpand_wildcards parameter to the indices.delete_data_stream and indices.get_data_stream APIsindices.delete_index_template, indices.exists_index_template, indices.get_index_template, indices.put_index_template, indices.simulate_index_template, and indices.simulate_template APIs moved from experimental to stableml.upgrade_job_snapshot APIexclude_generated parameter to ml.get_data_frame_analytics, ml.get_datafeeds, ml.get_jobs, and ml.get_trained_models APIsml.delete_data_frame_analytics, ml.delete_trained_model, ml.explain_data_frame_analytics, ml.get_data_frame_analytics, ml.get_data_frame_analytics_stats, ml.get_trained_models, ml.get_trained_models_stats, ml.put_data_frame_analytics, ml.put_trained_model, ml.start_data_frame_analytics, ml.stop_data_frame_analytics, ml.update_data_frame_analytics APIs from experimental to betarollup.rollup APIexclude_generated parameter to transform.get_transform APIwatcher.query_watches APIAdded ElasticsearchWarning which is raised when the Warning HTTP header is returned from Elasticsearch. ElasticsearchDeprecationWarning is now an alia…
X-Elastic-Client-Meta HTTP header and the meta_header parameter for controlling the header (#1473)ElasticsearchWarning which is raised when the Warning HTTP header is returned from Elasticsearch. ElasticsearchDeprecationWarning is now an alias for this warning type (#1495)async_search.status APIautoscaling.get_autoscaling_capacity experimental APIautoscaling.get_autoscaling_decision experimental APIcat.tasks API parameters node_id and parent_task to nodes and parent_task_idcluster.delete_component_template, cluster.exists_component_template, and cluster.get_component_template,
cluster.put_component_template APIs from experimental to stableeql.delete, eql.get, eql.search APIs from beta to stableindices.migrate_to_data_stream APIindices.promote_data_stream APIexpand_wildcards parameter to the indices.delete_data_stream and indices.get_data_stream APIsindices.delete_index_template, indices.exists_index_template, indices.get_index_template, indices.put_index_template, indices.simulate_index_template, and indices.simulate_template APIs moved from experimental to stableml.upgrade_job_snapshot APIexclude_generated parameter to ml.get_data_frame_analytics, ml.get_datafeeds, ml.get_jobs, and ml.get_trained_models APIsml.delete_data_frame_analytics, ml.delete_trained_model, ml.explain_data_frame_analytics, ml.get_data_frame_analytics, ml.get_data_frame_analytics_stats, ml.get_trained_models, ml.get_trained_models_stats, ml.put_data_frame_analytics, ml.put_trained_model, ml.start_data_frame_analytics, ml.stop_data_frame_analytics, ml.update_data_frame_analytics APIs from experimental to betarollup.rollup APIexclude_generated parameter to transform.get_transform APIwatcher.query_watches APIChanged deprecated collections.Mapping in favor of collections.abc.Mapping for Python 3.9
scroll response returned without a value for _shards.skipped (#1451)Connection.host property (#1460)collections.Mapping in favor of collections.abc.Mapping for Python 3.9 (#1443)Deprecated allow_no_jobs parameter of cat.ml_jobs and ml.close_job, ml.get_job_stats, ml.get_jobs, ml.get_overall_buckets methods in favor of allow_no…
_if_seq_no/_if_primary_term) to bulk helpers (#1387)_source with "_op_type": "update" bulk helpers (#1387)yarl dependency from [async] extra to avoid issue where pip would override aiohttp's pin of yarl. This was not a problem if you install with --use-feature=2020-resolver. Users should see no changes. (#1401)Connection.log_request_failure() call would receive the compressed HTTP body rather than uncompressed when an error is raised for RequestsHttpConnection (#1394)sniff_timeout was used instead of sniffer_timeout (#1431, contributed by @HarrySky)snapshot.clone methodrequire_alias parameter to multiple index APIsclose_point_in_time and open_point_in_time methodssecurity.clear_api_key_cache and security.grant_api_key methodsallow_no_jobs parameter of cat.ml_jobs and ml.close_job, ml.get_job_stats, ml.get_jobs, ml.get_overall_buckets methods in favor of allow_no_match parameterallow_no_datafeeds parameter of ml.get_datafeed_stats, ml.get_datafeeds, ml.stop_datafeed in favor of allow_no_match parameterinclude_model_definition parameter of ml.get_trained_models method in favor of include parameterAdded support for Elasticsearch 7.10 APIs
_if_seq_no/_if_primary_term) to bulk helpers (#1387)_source with "_op_type": "update" bulk helpers (#1387)yarl dependency from [async] extra to avoid issue where pip would override aiohttp's pin of yarl. This was not a problem if you install with --use-feature=2020-resolver. Users should see no changes. (#1401)Connection.log_request_failure() call would receive the compressed HTTP body rather than uncompressed when an error is raised for RequestsHttpConnection (#1394)Added support for Elasticsearch 7.10 APIs
_if_seq_no/_if_primary_term) to bulk helpers (#1387)_source with "_op_type": "update" bulk helpers (#1387)yarl dependency from [async] extra to avoid issue where pip would override aiohttp's pin of yarl. This was not a problem if you install with --use-feature=2020-resolver. Users should see no changes. (#1401)Connection.log_request_failure() call would receive the compressed HTTP body rather than uncompressed when an error is raised for RequestsHttpConnection (#1394)Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →