Status of testing Providers that were prepared on July 07, 2026

[vincbeck]

I have a kind request for all the contributors to the latest provider distributions release.
Could you please help us to test the RC versions of the providers?

The guidelines on how to test providers can be found in

Verify providers by contributors

Let us know in the comments, whether the issue is addressed.

These are providers that require testing as there were some substantial changes introduced:

Provider airbyte: 5.5.2rc1

• Cap airbyte-api below 1.0.0 to unblock provider tests (#69081): @shahar1
  Linked issues:
  • Linked Issue #68882: @aaronsteers

Provider amazon: 9.32.0rc1

• Add CRUD methods to DynamoDBHook (#69142): @gocoolp
  Linked issues:
  • Linked Issue #69050: @gocoolp
• Add Amazon Provider Neptune Analytics (#64274): @ellisms
  Linked issues:
  • Linked Issue #60975: @CincyBC
• Add team_name tags to Amazon executor metrics (#69072): @justinpakzad
  Linked issues:
  • Linked Issue #68996: @ferruzzi
• Add S3CopyPrefixOperator to copy all objects under a prefix (#68946): @bentorb
• Add the ability for SageMakerProcessingOperator to read JSON output files from S3. (#69002): @bhavya2109sharma
• Propagate AWS hook parameters through RedshiftClusterTrigger (#68925): @o-nikolas
  Linked issues:
  • Linked Issue #35278: @Taragolis
• Propagate verify and botocore_config through EC2StateSensorTrigger (#68921): @o-nikolas
  Linked issues:
  • Linked Issue #35278: @Taragolis
• Fix AWS Batch Executor compatibility with Airflow 3 in try_adopt_task_instances (#68027): @ambaena
  Linked issues:
  • Linked Issue #62192: @vincbeck
• prevent path traversal in s3 to sftp/ftp transfer destinations (#68984): @Samin061
• Fix SageMakerTransformOperator succeeding on a failed deferred job (#69042): @steveahnahn
  Linked issues:
  • Linked Issue #68927: @o-nikolas
• Close S3 download_file handle after download or failure (#69084): @itej13
  Linked issues:
  • Linked Issue #69077: @goransh-buh
• Make Amazon SageMaker triggers inherit AWS base classes (#68927): @o-nikolas
  Linked issues:
  • Linked Issue #35278: @Taragolis
• Fix batchoperator deferrable xcom links (#64745): @kakatur
• Standardize ECS TaskDoneTrigger on region_name and AWS hook parameters (#68923): @o-nikolas
  Linked issues:
  • Linked Issue #35278: @Taragolis
• Fix CloudWatch remote logging for ephemeral lifecycle executor (#68779): @jason810496
  Linked issues:
  • Linked Issue #66633: @korex-f
• Fix Cloudwatch remote logging crash API server if ResourceNotFoundException (#68781): @jason810496

Provider anthropic: 0.1.0rc1 🎉 New provider

Note: This is a new provider, so the PRs below are collected from the git history of the provider's sources rather than from a changelog diff. The list may therefore include bootstrap and development commits (e.g. release-preparation commits) that are not user-facing changes.

• Add Anthropic provider (apache-airflow-providers-anthropic) (#69003): @kaxil

Provider celery: 3.22.0rc1

• Honor json_logs config in Celery worker startup (#68916): @hanxdatadog
  Linked issues:
  • Linked Issue #68912: @hanxdatadog
• Fix Celery worker JSON logging config for teams (#69139): @henry3260
• Make PostgreSQL SQLAlchemy driver explicit (postgresql+psycopg2://) (#68314): @Dev-iL
  Linked issues:
  • Linked Issue #67801: @Dev-iL
• docs: clarify Redis maintenance for CeleryExecutor (#67393): @Will-thom
  Linked issues:
  • Linked Issue #55215: @gudata

Provider common.ai: 0.6.0rc1

• Add env_provider and Extra.env support to MCP stdio transport (#69225): @kaxil
• Replace deprecated pydantic-ai MCP classes with MCPToolset in common.ai (#69006): @kaxil
• Migrate common.ai provider to pydantic-ai 2.x and remove the <2 cap (#69358): @kaxil
  Linked issues:
  • Linked Issue #68933: @github-actions[bot]

Provider common.compat: 1.16.0rc1

• Added get_async_hook in common.compat provider (#69208): @dabla
• Fix RecursionError in common.compat hook lineage add_extra polyfill (#68735): @kaxil

Provider common.sql: 2.0.2rc1

• Fix DatabricksSqlHook crash on empty result with fetch_one handler (#69230): @moomindani
  Linked issues:
  • Linked Issue #39448: @KanikaAdik

Provider databricks: 7.17.0rc1

• Add durable execution to DatabricksRunNowOperator (#69174): @amoghrajesh
  Linked issues:
  • Linked Issue #68974: @amoghrajesh
• Accept BLOCKED and WAITING_FOR_RETRY Databricks run states (#69193): @amoghrajesh
• Add durable execution to DatabricksSubmitRunOperator (#68974): @amoghrajesh
• Add proxy support to Databricks connections (#68527): @Vamsi-klu
  Linked issues:
  • Linked Issue #32576: @sravanscb
• Surface databricks API errors as a separate exception (#69019): @amoghrajesh
• Add query tags: DatabricksSqlSensor, DatabricksPartitionSensor (#68704): @cruseakshay
  Linked issues:
  • Linked Issue #68582: @zhaorui2022
  • Linked Issue #66895: @SkastVnT
• Add OpenLineage Spark conf injection to DatabricksSubmitRunOperator (#67894): @rahul-madaan
• Include last error cause in Databricks API when retry is exhausted (#69238): @amoghrajesh
• Fix DatabricksSqlHook crash on empty result with fetch_one handler (#69230): @moomindani
  Linked issues:
  • Linked Issue #39448: @KanikaAdik
• Reconnect to blocked Databricks runs in SubmitRunOperator durable retries (#69195): @amoghrajesh
• Fix DatabricksWorkflowTaskGroup ignoring deps set before with block (#68924): @moomindani
  Linked issues:
  • Linked Issue #67369: @shubhgurav0590
  • Linked Issue #51598: @cshaddox

Provider edge3: 4.1.0rc1

• Add team_name tag to Edge executor sync metric (#69005): @shivaam
  Linked issues:
  • Linked Issue #68593: @ferruzzi
  • Linked Issue #68996: @ferruzzi

Provider elasticsearch: 6.8.0rc1

• Show running task logs in the UI with Elasticsearch remote logging (#69090): @aeroyorch
• ElasticsearchSQLHook: add chunked Polars DataFrame support via custom SQL reader (#68411): @SameerMesiah97
  Linked issues:
  • Linked Issue #66290: @SameerMesiah97
  • Linked Issue #66220: @SameerMesiah97

Provider exasol: 4.10.4rc1

• Fix DatabricksSqlHook crash on empty result with fetch_one handler (#69230): @moomindani
  Linked issues:
  • Linked Issue #39448: @KanikaAdik

Provider fab: 3.7.2rc1

• Fix DAG named "DAGs" colliding with the global DAGs permission resource (#69106): @potiuk

Provider ftp: 3.15.2rc1

• Close FTP file handles when a transfer raises (#69083): @itej13
  Linked issues:
  • Linked Issue #69077: @goransh-buh

Provider git: 0.4.1rc1

• Verify SSH host keys by default in the Git provider hook (#69103): @ephraimbuddy

Provider google: 22.2.2rc1

• Log SQL query on its own line in BigQueryInsertJobOperator (#69386): @AlejandroMorgante
  Linked issues:
  • Linked Issue #43953: @moiseenkov

Provider imap: 3.12.0rc1

• Add overwrite parameter to IMAP hook to prevent silent file overwrites #65870 (#68838): @kada2004
  Linked issues:
  • Linked Issue #65870: @vaefremov95

Provider microsoft.azure: 13.5.1rc1

• Fixed KiotaRequestAdapterHook _is_http_client_closed method wrongly returning True when transport was not opened yet (#69329): @dabla
• Invalidate cached RequestAdapter of KiotaRequestAdapterHook if session of AuthenticationProvider is closed (#69128): @dabla
• Fix msgraph/Power BI auth failure from empty allowed_hosts list (#69014): @davidnzhang
• Fix UnicodeDecodeError in DefaultResponseHandler when response content is binary (#68495): @davidnzhang

Provider microsoft.winrm: 3.14.4rc1

• Fix Microsoft WinRM provider import failure under lowest dependencies (#69333): @potiuk
  Linked issues:
  • Linked Issue #64643: @haseebmalik18

Provider neo4j: 3.12.0rc1

• Rename Neo4jOperator sql parameter to cypher (#68970): @eladkal
  Linked issues:
  • Linked Issue #39860: @ThomasThelen

Provider openai: 1.8.0rc1

• Deprecate OpenAI Assistants and Threads hook methods (#69071): @kaxil
  Linked issues:
  • Linked Issue #69068: @kaxil
  • Linked Issue #69070: @kaxil
• Add OpenAI Responses and Conversations support (#69070): @kaxil
  Linked issues:
  • Linked Issue #69068: @kaxil
• Add Workload Identity authentication to OpenAI provider (#69069): @kaxil
  Linked issues:
  • Linked Issue #69068: @kaxil
• Update OpenAI provider to OpenAI Python SDK 2.x (#69068): @kaxil

Provider openlineage: 2.19.0rc1

• Add unified emission_policy config for OpenLineage provider (#66992): @kacpermuda
• Add OpenLineage Spark conf injection to DatabricksSubmitRunOperator (#67894): @rahul-madaan

Provider opensearch: 1.11.0rc1

• Show running task logs in the UI with OpenSearch remote logging (#69407): @aeroyorch
  Linked issues:
  • Linked Issue #69090: @aeroyorch
• opensearch: generalize extra parameters (#65613): @DonHaul

Provider oracle: 4.6.2rc1

• Use oracledb AuthMode/Purity enums in Oracle hook connection config (#67689): @ed-kyu

Provider postgres: 7.0.0rc1

• Switch the default async Postgres driver from asyncpg to psycopg3 (provider) (#69089): @Dev-iL

Provider smtp: 3.0.2rc1

• Fixed "AttributeError: 'NoneType' object has no attribute 'close'" for SmtpHook (#62409): @TopCoder2K

Provider snowflake: 6.15.0rc1

• Add SnowparkContainerJobOperator (#68259): @justinpakzad
• Add SnowflakeCortexAgentHook (#68942): @SameerMesiah97
• Fix SnowflakeSqlApiOperator polling to avoid needless sleeps (#69450): @amoghrajesh
• Fix DatabricksSqlHook crash on empty result with fetch_one handler (#69230): @moomindani
  Linked issues:
  • Linked Issue #39448: @KanikaAdik
• SnowflakeHook: refactor static connection parameter construction (#68597): @SameerMesiah97
• SnowflakeHook: extract OAuth token lifecycle management into dedicated helper (#68549): @SameerMesiah97

Provider ssh: 5.0.4rc1

• Fix SSHRemoteJobOperator orphaning the remote job on cancellation (#68644): @kaxil

Provider standard: 1.16.0rc1

• Propagate partition_date to consumers of partitioned assets (#67285): @nathadfield
  Linked issues:
  • Linked Issue #68266: @Lee-W
  • Linked Issue #67239: @nathadfield

Provider tableau: 5.6.0rc1

• Add timeout and exponential backoff to Tableau refresh wait (#69479): @jayamanikharyono
• Import tableauserverclient lazily in TableauHook (#68971): @potiuk

All users involved in the PRs:
@AlejandroMorgante @bentorb @TopCoder2K @aeroyorch @nathadfield @jayamanikharyono @gocoolp @DonHaul @bhavya2109sharma @davidnzhang @cruseakshay @henry3260 @justinpakzad @shivaam @kada2004 @Samin061 @shahar1 @Will-thom @potiuk @amoghrajesh @Dev-iL @steveahnahn @Vamsi-klu @SameerMesiah97 @kacpermuda @rahul-madaan @itej13 @dabla @eladkal @jason810496 @ellisms @ephraimbuddy @ambaena @kaxil @kakatur @hanxdatadog @ed-kyu @moomindani @o-nikolas

[bhavya2109sharma]

Tested PR #69002 - apache-airflow-providers-amazon==9.32.0rc1 and it's working as expected.

[vaefremov95]

Hi!
Tested Imap 3.12.0rc1. Everything works fine.

[justinpakzad]

Confirming both #68259 and #69072 have been tested and are working as expected.

[moomindani]

Tested my PRs on the RC packages (installed from the 2026-07-06 dist artifacts, on Airflow 3.2.2):

• Fix DatabricksSqlHook crash on empty result with fetch_one handler #69230 (common-sql==2.0.2rc1, databricks==7.17.0rc1): validated against a live Databricks SQL warehouse — fetch_one on an empty result now returns None (previously TypeError), and non-empty / empty fetch_all, single-row, and multi-statement paths all pass with XCom-serializable results. For exasol==4.10.4rc1 and snowflake==6.15.0rc1 (annotation-only alignment from the same PR), hooks import cleanly and the run() return annotations include the None-able forms.
• Fix DatabricksWorkflowTaskGroup ignoring deps set before with block #68924 (databricks==7.17.0rc1): DatabricksWorkflowTaskGroup dependencies set with >> before the with block now reach the launch task (upstream) and leaf tasks (downstream); dependencies set after the block keep working.

Everything works as expected.

---

Drafted-by: Claude Code (Fable 5); reviewed by @moomindani before posting

[AlejandroMorgante]

Verified apache-airflow-providers-google==22.2.2rc1.

Tested: #69386

Environment:

• Breeze (Airflow main dev environment)
• Python 3.10
• Provider installed from the PyPI RC package (confirmed the operator module resolves to site-packages with version 22.2.2rc1, not local sources)
• Tested against a real GCP project (BigQuery, asia-east1 dataset)

Result:

• Ran the bq_value_check_location system test with a multiline SQL query in BigQueryInsertJobOperator — the full Dag passed (create dataset/table, insert job, value checks, teardown).
• The task log now shows the SQL on its own SQL query: line with natural newlines, right after the Executing: configuration line.
• The BigQuery job completed successfully, so no functional regression.

[SkastVnT]

Tested #68704 with apache-airflow-providers-databricks==7.17.0rc1 on a real Airflow + Databricks SQL Warehouse setup.
Query tags are visible in system.query.history and everything works as expected.

[steveahnahn]

Tested #69042 on amazon 9.32.0rc1 and confirmed SageMakerTransformOperator.execute_complete now raises RuntimeError on a failed deferred transform job instead of reporting success. Working as expected.

[amoghrajesh]

Installed the release bits and tested out my changes -

• Fix SnowflakeSqlApiOperator polling to avoid needless sleeps #69450
• Add durable execution to DatabricksRunNowOperator #69174
• Add durable execution to DatabricksSubmitRunOperator #68974
• Accept BLOCKED and WAITING_FOR_RETRY Databricks run states #69193
• Add durable execution to DatabricksSubmitRunOperator #68974
• Surface databricks API errors as a separate exception #69019
• Include last error cause in Databricks API when retry is exhausted #69238
• Reconnect to blocked Databricks runs in SubmitRunOperator durable retries #69195

Changes look great.

[Vamsi-klu]

Tested apache-airflow-providers-databricks==7.17.0rc1 against Airflow 3.1.3 in Breeze using a live Databricks workspace.

Validation covered an authenticated DatabricksHook call to GET /api/2.0/workspace/list through a local HTTP CONNECT proxy. Both sync and async hook paths returned successful API responses and the proxy captured CONNECT tunnels to the workspace host, so the proxy support added in #68527 looks good for this RC.

I also checked clusters/list; that request reached Databricks through the proxy but returned 403 because the temporary token did not include the clusters scope, so I used the workspace-scoped endpoint for the final authenticated validation.

---

Drafted-by: Codex (GPT-5); reviewed by @Vamsi-klu before posting

[Samin061]

Tested #68984 on apache-airflow-providers-amazon==9.32.0rc1 (installed from the PyPI RC, confirmed the module resolves to site-packages at 9.32.0rc1, not local sources).

The path-traversal guard ships in the RC: both S3ToSFTPOperator and S3ToFTPOperator call validate_destination_path before writing. Attacker-controllable S3 key suffixes with .. segments (e.g. /srv/sftp/incoming/../../../../etc/cron.d/evil, /var/ftp/out/../../etc/passwd) now raise ValueError instead of escaping the configured sftp_path/ftp_path, while benign nested suffixes like sub/dir/report.csv are still allowed. Working as expected.

[molcay]

Google provider seems to work as expected.

[gocoolp]

Tested PR #69142 — Add CRUD methods to DynamoDBHook on apache-airflow-providers-amazon==9.32.0rc1.

Environment:

• Python 3.13
• apache-airflow-providers-amazon==9.32.0rc1
• moto (mocked DynamoDB)

Methods validated:

• put_item ✓
• get_item ✓ (including None return for missing item)
• update_item ✓
• delete_item ✓
• query ✓

All CRUD methods are working as expected.