Fix unclosed aiohttp ClientSession in AzureDataFactoryAsyncHook

[cruseakshay]

Closes: #55724

Fix unclosed aiohttp.ClientSession and TCPConnector warnings when using AzureDataFactoryRunPipelineOperator with deferrable=True.

The AzureDataFactoryAsyncHook creates an AsyncDataFactoryManagementClient but never closes it, causing resource leak errors in logs regardless of pipeline success/failure.

Changes

• Add __aenter__, __aexit__, and close() methods to AzureDataFactoryAsyncHook
• Update triggers to use async with hook: for proper cleanup
• Add unit tests for context manager behavior

Follows the same pattern as #52119 (Databricks fix).

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Claude following the guidelines

---

[SameerMesiah97]

Looks good to me. Turning the async hook into a context manager and updating the trigger call sites fixes the aiohttp client leak cleanly. Tests look solid as well.

Not a blocker but I would suggest double checking if there any places outside this file where this hook is being called so that the leak does not persist after your fix.

[cruseakshay]

@SameerMesiah97 Double checked other places, looks good.

[dabla]

2 small nits but looking good!

[dabla]

Looking good to me.

[dabla]

@cruseakshay Did you actually tests this changes and can you confirm this works?

[dabla]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

[cruseakshay]

@cruseakshay Did you actually tests this changes and can you confirm this works?

Yes, here are trail logs from latest run

[2026-01-23T04:53:13.235944Z] INFO - Sleeping for 60. The pipeline state is Queued.
[2026-01-23T04:54:13.245821Z] INFO - Request URL: 'https://management.azure.com/subscriptions/REDACTED/resourceGroups/airflow-adf-test-rg/providers/Microsoft.DataFactory/factories/airflow-adf-test/pipelineruns/REDACTED?api-version=REDACTED'
Request method: 'GET'
Request headers:
'Accept': 'application/json'
'x-ms-client-request-id': '9001a594-f817-11f0-b40d-b64b850643a2'
'User-Agent': 'azsdk-python-mgmt-datafactory/9.2.0 Python/3.12.12 (Linux-6.12.54-linuxkit-aarch64-with-glibc2.36)'
'Authorization': 'REDACTED'
No body was attached to the request
[2026-01-23T04:54:13.973572Z] INFO - Response status: 200
Response headers:
'Cache-Control': 'no-cache'
'Pragma': 'no-cache'
'Content-Length': '816'
'Content-Type': 'application/json; charset=utf-8'
'Expires': '-1'
'x-ms-correlation-request-id': 'REDACTED'
'x-ms-operation-identifier': 'REDACTED'
'x-ms-ratelimit-remaining-subscription-reads': '249'
'x-ms-ratelimit-remaining-subscription-global-reads': 'REDACTED'
'x-ms-request-id': '59f3995c-910c-4abe-952a-5e3e8fa3bff3'
'x-ms-routing-request-id': 'REDACTED'
'Strict-Transport-Security': 'REDACTED'
'X-Content-Type-Options': 'REDACTED'
'X-Cache': 'REDACTED'
'X-MSEdge-Ref': 'Ref A: 1C93C26BC61B478A8D2F48098F9A042C Ref B: MAA201060513029 Ref C: 2026-01-23T04:54:13Z'
'Date': 'Fri, 23 Jan 2026 04:54:13 GMT'
[2026-01-23T04:54:13.974417Z] INFO - Trigger fired eventname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) result=TriggerEvent<{'status': 'success', 'message': 'The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.', 'run_id': '66bb30ba-f817-11f0-bec5-16db0266592f'}> [2026-01-23T04:54:13.975233Z] INFO - trigger completedname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) [2026-01-23T04:54:15.380187Z] INFO - DAG bundles loaded: dags-folder
[2026-01-23T04:54:15.380487Z] INFO - Filling up the DagBag from /usr/local/airflow/dags/test_adf_trigger_fix.py
[2026-01-23T04:54:15.488568Z] INFO - TaskInstance Details: dag_id=test_adf_trigger_fix task_id=run_pipeline_deferrable dagrun_id=manual__2026-01-23T04:53:02.479036+00:00 map_index=-1 run_start_date=2026-01-23T04:54:15.366708Z try_number=1 op_classpath=["airflow.providers.microsoft.azure.operators.data_factory.AzureDataFactoryRunPipelineOperator"] [2026-01-23T04:54:15.488765Z] INFO - The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.
[2026-01-23T04:54:15.508872Z] WARNING - Connection schemes (type: azure_data_factory) shall not contain '_' according to RFC3986.

[cruseakshay]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

hook-level test that covers the context manager behavior: AzureDataFactoryAsyncHook has test_context_manager_calls_close that explicitly verifies async with hook: calls close(). Haven't seen any trigger tests explicitly assert __aenter__/__aexit__ calls.

[dabla]

@cruseakshay Thanks for the PR and your patience, good job!

[boring-cyborg]

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

[cruseakshay]

Closes: #55724

Fix unclosed aiohttp.ClientSession and TCPConnector warnings when using AzureDataFactoryRunPipelineOperator with deferrable=True.

The AzureDataFactoryAsyncHook creates an AsyncDataFactoryManagementClient but never closes it, causing resource leak errors in logs regardless of pipeline success/failure.

Changes

• Add __aenter__, __aexit__, and close() methods to AzureDataFactoryAsyncHook
• Update triggers to use async with hook: for proper cleanup
• Add unit tests for context manager behavior

Follows the same pattern as #52119 (Databricks fix).

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Claude following the guidelines

---

[SameerMesiah97]

Looks good to me. Turning the async hook into a context manager and updating the trigger call sites fixes the aiohttp client leak cleanly. Tests look solid as well.

Not a blocker but I would suggest double checking if there any places outside this file where this hook is being called so that the leak does not persist after your fix.

[cruseakshay]

@SameerMesiah97 Double checked other places, looks good.

[dabla]

2 small nits but looking good!

[dabla]

Looking good to me.

[dabla]

@cruseakshay Did you actually tests this changes and can you confirm this works?

[dabla]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

[cruseakshay]

@cruseakshay Did you actually tests this changes and can you confirm this works?

Yes, here are trail logs from latest run

[2026-01-23T04:53:13.235944Z] INFO - Sleeping for 60. The pipeline state is Queued.
[2026-01-23T04:54:13.245821Z] INFO - Request URL: 'https://management.azure.com/subscriptions/REDACTED/resourceGroups/airflow-adf-test-rg/providers/Microsoft.DataFactory/factories/airflow-adf-test/pipelineruns/REDACTED?api-version=REDACTED'
Request method: 'GET'
Request headers:
'Accept': 'application/json'
'x-ms-client-request-id': '9001a594-f817-11f0-b40d-b64b850643a2'
'User-Agent': 'azsdk-python-mgmt-datafactory/9.2.0 Python/3.12.12 (Linux-6.12.54-linuxkit-aarch64-with-glibc2.36)'
'Authorization': 'REDACTED'
No body was attached to the request
[2026-01-23T04:54:13.973572Z] INFO - Response status: 200
Response headers:
'Cache-Control': 'no-cache'
'Pragma': 'no-cache'
'Content-Length': '816'
'Content-Type': 'application/json; charset=utf-8'
'Expires': '-1'
'x-ms-correlation-request-id': 'REDACTED'
'x-ms-operation-identifier': 'REDACTED'
'x-ms-ratelimit-remaining-subscription-reads': '249'
'x-ms-ratelimit-remaining-subscription-global-reads': 'REDACTED'
'x-ms-request-id': '59f3995c-910c-4abe-952a-5e3e8fa3bff3'
'x-ms-routing-request-id': 'REDACTED'
'Strict-Transport-Security': 'REDACTED'
'X-Content-Type-Options': 'REDACTED'
'X-Cache': 'REDACTED'
'X-MSEdge-Ref': 'Ref A: 1C93C26BC61B478A8D2F48098F9A042C Ref B: MAA201060513029 Ref C: 2026-01-23T04:54:13Z'
'Date': 'Fri, 23 Jan 2026 04:54:13 GMT'
[2026-01-23T04:54:13.974417Z] INFO - Trigger fired eventname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) result=TriggerEvent<{'status': 'success', 'message': 'The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.', 'run_id': '66bb30ba-f817-11f0-bec5-16db0266592f'}> [2026-01-23T04:54:13.975233Z] INFO - trigger completedname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) [2026-01-23T04:54:15.380187Z] INFO - DAG bundles loaded: dags-folder
[2026-01-23T04:54:15.380487Z] INFO - Filling up the DagBag from /usr/local/airflow/dags/test_adf_trigger_fix.py
[2026-01-23T04:54:15.488568Z] INFO - TaskInstance Details: dag_id=test_adf_trigger_fix task_id=run_pipeline_deferrable dagrun_id=manual__2026-01-23T04:53:02.479036+00:00 map_index=-1 run_start_date=2026-01-23T04:54:15.366708Z try_number=1 op_classpath=["airflow.providers.microsoft.azure.operators.data_factory.AzureDataFactoryRunPipelineOperator"] [2026-01-23T04:54:15.488765Z] INFO - The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.
[2026-01-23T04:54:15.508872Z] WARNING - Connection schemes (type: azure_data_factory) shall not contain '_' according to RFC3986.

[cruseakshay]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

hook-level test that covers the context manager behavior: AzureDataFactoryAsyncHook has test_context_manager_calls_close that explicitly verifies async with hook: calls close(). Haven't seen any trigger tests explicitly assert __aenter__/__aexit__ calls.

[dabla]

@cruseakshay Thanks for the PR and your patience, good job!

[boring-cyborg]

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

[cruseakshay]

Closes: #55724

Fix unclosed aiohttp.ClientSession and TCPConnector warnings when using AzureDataFactoryRunPipelineOperator with deferrable=True.

The AzureDataFactoryAsyncHook creates an AsyncDataFactoryManagementClient but never closes it, causing resource leak errors in logs regardless of pipeline success/failure.

Changes

• Add __aenter__, __aexit__, and close() methods to AzureDataFactoryAsyncHook
• Update triggers to use async with hook: for proper cleanup
• Add unit tests for context manager behavior

Follows the same pattern as #52119 (Databricks fix).

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Claude following the guidelines

---

[SameerMesiah97]

Looks good to me. Turning the async hook into a context manager and updating the trigger call sites fixes the aiohttp client leak cleanly. Tests look solid as well.

Not a blocker but I would suggest double checking if there any places outside this file where this hook is being called so that the leak does not persist after your fix.

[cruseakshay]

@SameerMesiah97 Double checked other places, looks good.

[dabla]

2 small nits but looking good!

[dabla]

Looking good to me.

[dabla]

@cruseakshay Did you actually tests this changes and can you confirm this works?

[dabla]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

[cruseakshay]

@cruseakshay Did you actually tests this changes and can you confirm this works?

Yes, here are trail logs from latest run

[2026-01-23T04:53:13.235944Z] INFO - Sleeping for 60. The pipeline state is Queued.
[2026-01-23T04:54:13.245821Z] INFO - Request URL: 'https://management.azure.com/subscriptions/REDACTED/resourceGroups/airflow-adf-test-rg/providers/Microsoft.DataFactory/factories/airflow-adf-test/pipelineruns/REDACTED?api-version=REDACTED'
Request method: 'GET'
Request headers:
'Accept': 'application/json'
'x-ms-client-request-id': '9001a594-f817-11f0-b40d-b64b850643a2'
'User-Agent': 'azsdk-python-mgmt-datafactory/9.2.0 Python/3.12.12 (Linux-6.12.54-linuxkit-aarch64-with-glibc2.36)'
'Authorization': 'REDACTED'
No body was attached to the request
[2026-01-23T04:54:13.973572Z] INFO - Response status: 200
Response headers:
'Cache-Control': 'no-cache'
'Pragma': 'no-cache'
'Content-Length': '816'
'Content-Type': 'application/json; charset=utf-8'
'Expires': '-1'
'x-ms-correlation-request-id': 'REDACTED'
'x-ms-operation-identifier': 'REDACTED'
'x-ms-ratelimit-remaining-subscription-reads': '249'
'x-ms-ratelimit-remaining-subscription-global-reads': 'REDACTED'
'x-ms-request-id': '59f3995c-910c-4abe-952a-5e3e8fa3bff3'
'x-ms-routing-request-id': 'REDACTED'
'Strict-Transport-Security': 'REDACTED'
'X-Content-Type-Options': 'REDACTED'
'X-Cache': 'REDACTED'
'X-MSEdge-Ref': 'Ref A: 1C93C26BC61B478A8D2F48098F9A042C Ref B: MAA201060513029 Ref C: 2026-01-23T04:54:13Z'
'Date': 'Fri, 23 Jan 2026 04:54:13 GMT'
[2026-01-23T04:54:13.974417Z] INFO - Trigger fired eventname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) result=TriggerEvent<{'status': 'success', 'message': 'The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.', 'run_id': '66bb30ba-f817-11f0-bec5-16db0266592f'}> [2026-01-23T04:54:13.975233Z] INFO - trigger completedname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) [2026-01-23T04:54:15.380187Z] INFO - DAG bundles loaded: dags-folder
[2026-01-23T04:54:15.380487Z] INFO - Filling up the DagBag from /usr/local/airflow/dags/test_adf_trigger_fix.py
[2026-01-23T04:54:15.488568Z] INFO - TaskInstance Details: dag_id=test_adf_trigger_fix task_id=run_pipeline_deferrable dagrun_id=manual__2026-01-23T04:53:02.479036+00:00 map_index=-1 run_start_date=2026-01-23T04:54:15.366708Z try_number=1 op_classpath=["airflow.providers.microsoft.azure.operators.data_factory.AzureDataFactoryRunPipelineOperator"] [2026-01-23T04:54:15.488765Z] INFO - The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.
[2026-01-23T04:54:15.508872Z] WARNING - Connection schemes (type: azure_data_factory) shall not contain '_' according to RFC3986.

[cruseakshay]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

hook-level test that covers the context manager behavior: AzureDataFactoryAsyncHook has test_context_manager_calls_close that explicitly verifies async with hook: calls close(). Haven't seen any trigger tests explicitly assert __aenter__/__aexit__ calls.

[dabla]

@cruseakshay Thanks for the PR and your patience, good job!

[boring-cyborg]

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

[cruseakshay]

Closes: #55724

Fix unclosed aiohttp.ClientSession and TCPConnector warnings when using AzureDataFactoryRunPipelineOperator with deferrable=True.

The AzureDataFactoryAsyncHook creates an AsyncDataFactoryManagementClient but never closes it, causing resource leak errors in logs regardless of pipeline success/failure.

Changes

• Add __aenter__, __aexit__, and close() methods to AzureDataFactoryAsyncHook
• Update triggers to use async with hook: for proper cleanup
• Add unit tests for context manager behavior

Follows the same pattern as #52119 (Databricks fix).

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Claude following the guidelines

---

[SameerMesiah97]

Looks good to me. Turning the async hook into a context manager and updating the trigger call sites fixes the aiohttp client leak cleanly. Tests look solid as well.

Not a blocker but I would suggest double checking if there any places outside this file where this hook is being called so that the leak does not persist after your fix.

[cruseakshay]

@SameerMesiah97 Double checked other places, looks good.

[dabla]

2 small nits but looking good!

[dabla]

Looking good to me.

[dabla]

@cruseakshay Did you actually tests this changes and can you confirm this works?

[dabla]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

[cruseakshay]

@cruseakshay Did you actually tests this changes and can you confirm this works?

Yes, here are trail logs from latest run

[2026-01-23T04:53:13.235944Z] INFO - Sleeping for 60. The pipeline state is Queued.
[2026-01-23T04:54:13.245821Z] INFO - Request URL: 'https://management.azure.com/subscriptions/REDACTED/resourceGroups/airflow-adf-test-rg/providers/Microsoft.DataFactory/factories/airflow-adf-test/pipelineruns/REDACTED?api-version=REDACTED'
Request method: 'GET'
Request headers:
'Accept': 'application/json'
'x-ms-client-request-id': '9001a594-f817-11f0-b40d-b64b850643a2'
'User-Agent': 'azsdk-python-mgmt-datafactory/9.2.0 Python/3.12.12 (Linux-6.12.54-linuxkit-aarch64-with-glibc2.36)'
'Authorization': 'REDACTED'
No body was attached to the request
[2026-01-23T04:54:13.973572Z] INFO - Response status: 200
Response headers:
'Cache-Control': 'no-cache'
'Pragma': 'no-cache'
'Content-Length': '816'
'Content-Type': 'application/json; charset=utf-8'
'Expires': '-1'
'x-ms-correlation-request-id': 'REDACTED'
'x-ms-operation-identifier': 'REDACTED'
'x-ms-ratelimit-remaining-subscription-reads': '249'
'x-ms-ratelimit-remaining-subscription-global-reads': 'REDACTED'
'x-ms-request-id': '59f3995c-910c-4abe-952a-5e3e8fa3bff3'
'x-ms-routing-request-id': 'REDACTED'
'Strict-Transport-Security': 'REDACTED'
'X-Content-Type-Options': 'REDACTED'
'X-Cache': 'REDACTED'
'X-MSEdge-Ref': 'Ref A: 1C93C26BC61B478A8D2F48098F9A042C Ref B: MAA201060513029 Ref C: 2026-01-23T04:54:13Z'
'Date': 'Fri, 23 Jan 2026 04:54:13 GMT'
[2026-01-23T04:54:13.974417Z] INFO - Trigger fired eventname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) result=TriggerEvent<{'status': 'success', 'message': 'The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.', 'run_id': '66bb30ba-f817-11f0-bec5-16db0266592f'}> [2026-01-23T04:54:13.975233Z] INFO - trigger completedname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) [2026-01-23T04:54:15.380187Z] INFO - DAG bundles loaded: dags-folder
[2026-01-23T04:54:15.380487Z] INFO - Filling up the DagBag from /usr/local/airflow/dags/test_adf_trigger_fix.py
[2026-01-23T04:54:15.488568Z] INFO - TaskInstance Details: dag_id=test_adf_trigger_fix task_id=run_pipeline_deferrable dagrun_id=manual__2026-01-23T04:53:02.479036+00:00 map_index=-1 run_start_date=2026-01-23T04:54:15.366708Z try_number=1 op_classpath=["airflow.providers.microsoft.azure.operators.data_factory.AzureDataFactoryRunPipelineOperator"] [2026-01-23T04:54:15.488765Z] INFO - The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.
[2026-01-23T04:54:15.508872Z] WARNING - Connection schemes (type: azure_data_factory) shall not contain '_' according to RFC3986.

[cruseakshay]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

hook-level test that covers the context manager behavior: AzureDataFactoryAsyncHook has test_context_manager_calls_close that explicitly verifies async with hook: calls close(). Haven't seen any trigger tests explicitly assert __aenter__/__aexit__ calls.

[dabla]

@cruseakshay Thanks for the PR and your patience, good job!

[boring-cyborg]

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

[cruseakshay]

Closes: #55724

Fix unclosed aiohttp.ClientSession and TCPConnector warnings when using AzureDataFactoryRunPipelineOperator with deferrable=True.

The AzureDataFactoryAsyncHook creates an AsyncDataFactoryManagementClient but never closes it, causing resource leak errors in logs regardless of pipeline success/failure.

Changes

• Add __aenter__, __aexit__, and close() methods to AzureDataFactoryAsyncHook
• Update triggers to use async with hook: for proper cleanup
• Add unit tests for context manager behavior

Follows the same pattern as #52119 (Databricks fix).

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Claude following the guidelines

---

[SameerMesiah97]

Looks good to me. Turning the async hook into a context manager and updating the trigger call sites fixes the aiohttp client leak cleanly. Tests look solid as well.

Not a blocker but I would suggest double checking if there any places outside this file where this hook is being called so that the leak does not persist after your fix.

[cruseakshay]

@SameerMesiah97 Double checked other places, looks good.

[dabla]

2 small nits but looking good!

[dabla]

Looking good to me.

[dabla]

@cruseakshay Did you actually tests this changes and can you confirm this works?

[dabla]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

[cruseakshay]

@cruseakshay Did you actually tests this changes and can you confirm this works?

Yes, here are trail logs from latest run

[2026-01-23T04:53:13.235944Z] INFO - Sleeping for 60. The pipeline state is Queued.
[2026-01-23T04:54:13.245821Z] INFO - Request URL: 'https://management.azure.com/subscriptions/REDACTED/resourceGroups/airflow-adf-test-rg/providers/Microsoft.DataFactory/factories/airflow-adf-test/pipelineruns/REDACTED?api-version=REDACTED'
Request method: 'GET'
Request headers:
'Accept': 'application/json'
'x-ms-client-request-id': '9001a594-f817-11f0-b40d-b64b850643a2'
'User-Agent': 'azsdk-python-mgmt-datafactory/9.2.0 Python/3.12.12 (Linux-6.12.54-linuxkit-aarch64-with-glibc2.36)'
'Authorization': 'REDACTED'
No body was attached to the request
[2026-01-23T04:54:13.973572Z] INFO - Response status: 200
Response headers:
'Cache-Control': 'no-cache'
'Pragma': 'no-cache'
'Content-Length': '816'
'Content-Type': 'application/json; charset=utf-8'
'Expires': '-1'
'x-ms-correlation-request-id': 'REDACTED'
'x-ms-operation-identifier': 'REDACTED'
'x-ms-ratelimit-remaining-subscription-reads': '249'
'x-ms-ratelimit-remaining-subscription-global-reads': 'REDACTED'
'x-ms-request-id': '59f3995c-910c-4abe-952a-5e3e8fa3bff3'
'x-ms-routing-request-id': 'REDACTED'
'Strict-Transport-Security': 'REDACTED'
'X-Content-Type-Options': 'REDACTED'
'X-Cache': 'REDACTED'
'X-MSEdge-Ref': 'Ref A: 1C93C26BC61B478A8D2F48098F9A042C Ref B: MAA201060513029 Ref C: 2026-01-23T04:54:13Z'
'Date': 'Fri, 23 Jan 2026 04:54:13 GMT'
[2026-01-23T04:54:13.974417Z] INFO - Trigger fired eventname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) result=TriggerEvent<{'status': 'success', 'message': 'The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.', 'run_id': '66bb30ba-f817-11f0-bec5-16db0266592f'}> [2026-01-23T04:54:13.975233Z] INFO - trigger completedname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) [2026-01-23T04:54:15.380187Z] INFO - DAG bundles loaded: dags-folder
[2026-01-23T04:54:15.380487Z] INFO - Filling up the DagBag from /usr/local/airflow/dags/test_adf_trigger_fix.py
[2026-01-23T04:54:15.488568Z] INFO - TaskInstance Details: dag_id=test_adf_trigger_fix task_id=run_pipeline_deferrable dagrun_id=manual__2026-01-23T04:53:02.479036+00:00 map_index=-1 run_start_date=2026-01-23T04:54:15.366708Z try_number=1 op_classpath=["airflow.providers.microsoft.azure.operators.data_factory.AzureDataFactoryRunPipelineOperator"] [2026-01-23T04:54:15.488765Z] INFO - The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.
[2026-01-23T04:54:15.508872Z] WARNING - Connection schemes (type: azure_data_factory) shall not contain '_' according to RFC3986.

[cruseakshay]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

hook-level test that covers the context manager behavior: AzureDataFactoryAsyncHook has test_context_manager_calls_close that explicitly verifies async with hook: calls close(). Haven't seen any trigger tests explicitly assert __aenter__/__aexit__ calls.

[dabla]

@cruseakshay Thanks for the PR and your patience, good job!

[boring-cyborg]

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

[cruseakshay]

Closes: #55724

Fix unclosed aiohttp.ClientSession and TCPConnector warnings when using AzureDataFactoryRunPipelineOperator with deferrable=True.

The AzureDataFactoryAsyncHook creates an AsyncDataFactoryManagementClient but never closes it, causing resource leak errors in logs regardless of pipeline success/failure.

Changes

• Add __aenter__, __aexit__, and close() methods to AzureDataFactoryAsyncHook
• Update triggers to use async with hook: for proper cleanup
• Add unit tests for context manager behavior

Follows the same pattern as #52119 (Databricks fix).

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Claude following the guidelines

---

[SameerMesiah97]

Looks good to me. Turning the async hook into a context manager and updating the trigger call sites fixes the aiohttp client leak cleanly. Tests look solid as well.

Not a blocker but I would suggest double checking if there any places outside this file where this hook is being called so that the leak does not persist after your fix.

[cruseakshay]

@SameerMesiah97 Double checked other places, looks good.

[dabla]

2 small nits but looking good!

[dabla]

Looking good to me.

[dabla]

@cruseakshay Did you actually tests this changes and can you confirm this works?

[dabla]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

[cruseakshay]

@cruseakshay Did you actually tests this changes and can you confirm this works?

Yes, here are trail logs from latest run

[2026-01-23T04:53:13.235944Z] INFO - Sleeping for 60. The pipeline state is Queued.
[2026-01-23T04:54:13.245821Z] INFO - Request URL: 'https://management.azure.com/subscriptions/REDACTED/resourceGroups/airflow-adf-test-rg/providers/Microsoft.DataFactory/factories/airflow-adf-test/pipelineruns/REDACTED?api-version=REDACTED'
Request method: 'GET'
Request headers:
'Accept': 'application/json'
'x-ms-client-request-id': '9001a594-f817-11f0-b40d-b64b850643a2'
'User-Agent': 'azsdk-python-mgmt-datafactory/9.2.0 Python/3.12.12 (Linux-6.12.54-linuxkit-aarch64-with-glibc2.36)'
'Authorization': 'REDACTED'
No body was attached to the request
[2026-01-23T04:54:13.973572Z] INFO - Response status: 200
Response headers:
'Cache-Control': 'no-cache'
'Pragma': 'no-cache'
'Content-Length': '816'
'Content-Type': 'application/json; charset=utf-8'
'Expires': '-1'
'x-ms-correlation-request-id': 'REDACTED'
'x-ms-operation-identifier': 'REDACTED'
'x-ms-ratelimit-remaining-subscription-reads': '249'
'x-ms-ratelimit-remaining-subscription-global-reads': 'REDACTED'
'x-ms-request-id': '59f3995c-910c-4abe-952a-5e3e8fa3bff3'
'x-ms-routing-request-id': 'REDACTED'
'Strict-Transport-Security': 'REDACTED'
'X-Content-Type-Options': 'REDACTED'
'X-Cache': 'REDACTED'
'X-MSEdge-Ref': 'Ref A: 1C93C26BC61B478A8D2F48098F9A042C Ref B: MAA201060513029 Ref C: 2026-01-23T04:54:13Z'
'Date': 'Fri, 23 Jan 2026 04:54:13 GMT'
[2026-01-23T04:54:13.974417Z] INFO - Trigger fired eventname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) result=TriggerEvent<{'status': 'success', 'message': 'The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.', 'run_id': '66bb30ba-f817-11f0-bec5-16db0266592f'}> [2026-01-23T04:54:13.975233Z] INFO - trigger completedname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) [2026-01-23T04:54:15.380187Z] INFO - DAG bundles loaded: dags-folder
[2026-01-23T04:54:15.380487Z] INFO - Filling up the DagBag from /usr/local/airflow/dags/test_adf_trigger_fix.py
[2026-01-23T04:54:15.488568Z] INFO - TaskInstance Details: dag_id=test_adf_trigger_fix task_id=run_pipeline_deferrable dagrun_id=manual__2026-01-23T04:53:02.479036+00:00 map_index=-1 run_start_date=2026-01-23T04:54:15.366708Z try_number=1 op_classpath=["airflow.providers.microsoft.azure.operators.data_factory.AzureDataFactoryRunPipelineOperator"] [2026-01-23T04:54:15.488765Z] INFO - The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.
[2026-01-23T04:54:15.508872Z] WARNING - Connection schemes (type: azure_data_factory) shall not contain '_' according to RFC3986.

[cruseakshay]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

hook-level test that covers the context manager behavior: AzureDataFactoryAsyncHook has test_context_manager_calls_close that explicitly verifies async with hook: calls close(). Haven't seen any trigger tests explicitly assert __aenter__/__aexit__ calls.

[dabla]

@cruseakshay Thanks for the PR and your patience, good job!

[boring-cyborg]

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

[cruseakshay]

Closes: #55724

Fix unclosed aiohttp.ClientSession and TCPConnector warnings when using AzureDataFactoryRunPipelineOperator with deferrable=True.

The AzureDataFactoryAsyncHook creates an AsyncDataFactoryManagementClient but never closes it, causing resource leak errors in logs regardless of pipeline success/failure.

Changes

• Add __aenter__, __aexit__, and close() methods to AzureDataFactoryAsyncHook
• Update triggers to use async with hook: for proper cleanup
• Add unit tests for context manager behavior

Follows the same pattern as #52119 (Databricks fix).

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Claude following the guidelines

---

[SameerMesiah97]

Looks good to me. Turning the async hook into a context manager and updating the trigger call sites fixes the aiohttp client leak cleanly. Tests look solid as well.

Not a blocker but I would suggest double checking if there any places outside this file where this hook is being called so that the leak does not persist after your fix.

[cruseakshay]

@SameerMesiah97 Double checked other places, looks good.

[dabla]

2 small nits but looking good!

[dabla]

Looking good to me.

[dabla]

@cruseakshay Did you actually tests this changes and can you confirm this works?

[dabla]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

[cruseakshay]

@cruseakshay Did you actually tests this changes and can you confirm this works?

Yes, here are trail logs from latest run

[2026-01-23T04:53:13.235944Z] INFO - Sleeping for 60. The pipeline state is Queued.
[2026-01-23T04:54:13.245821Z] INFO - Request URL: 'https://management.azure.com/subscriptions/REDACTED/resourceGroups/airflow-adf-test-rg/providers/Microsoft.DataFactory/factories/airflow-adf-test/pipelineruns/REDACTED?api-version=REDACTED'
Request method: 'GET'
Request headers:
'Accept': 'application/json'
'x-ms-client-request-id': '9001a594-f817-11f0-b40d-b64b850643a2'
'User-Agent': 'azsdk-python-mgmt-datafactory/9.2.0 Python/3.12.12 (Linux-6.12.54-linuxkit-aarch64-with-glibc2.36)'
'Authorization': 'REDACTED'
No body was attached to the request
[2026-01-23T04:54:13.973572Z] INFO - Response status: 200
Response headers:
'Cache-Control': 'no-cache'
'Pragma': 'no-cache'
'Content-Length': '816'
'Content-Type': 'application/json; charset=utf-8'
'Expires': '-1'
'x-ms-correlation-request-id': 'REDACTED'
'x-ms-operation-identifier': 'REDACTED'
'x-ms-ratelimit-remaining-subscription-reads': '249'
'x-ms-ratelimit-remaining-subscription-global-reads': 'REDACTED'
'x-ms-request-id': '59f3995c-910c-4abe-952a-5e3e8fa3bff3'
'x-ms-routing-request-id': 'REDACTED'
'Strict-Transport-Security': 'REDACTED'
'X-Content-Type-Options': 'REDACTED'
'X-Cache': 'REDACTED'
'X-MSEdge-Ref': 'Ref A: 1C93C26BC61B478A8D2F48098F9A042C Ref B: MAA201060513029 Ref C: 2026-01-23T04:54:13Z'
'Date': 'Fri, 23 Jan 2026 04:54:13 GMT'
[2026-01-23T04:54:13.974417Z] INFO - Trigger fired eventname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) result=TriggerEvent<{'status': 'success', 'message': 'The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.', 'run_id': '66bb30ba-f817-11f0-bec5-16db0266592f'}> [2026-01-23T04:54:13.975233Z] INFO - trigger completedname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) [2026-01-23T04:54:15.380187Z] INFO - DAG bundles loaded: dags-folder
[2026-01-23T04:54:15.380487Z] INFO - Filling up the DagBag from /usr/local/airflow/dags/test_adf_trigger_fix.py
[2026-01-23T04:54:15.488568Z] INFO - TaskInstance Details: dag_id=test_adf_trigger_fix task_id=run_pipeline_deferrable dagrun_id=manual__2026-01-23T04:53:02.479036+00:00 map_index=-1 run_start_date=2026-01-23T04:54:15.366708Z try_number=1 op_classpath=["airflow.providers.microsoft.azure.operators.data_factory.AzureDataFactoryRunPipelineOperator"] [2026-01-23T04:54:15.488765Z] INFO - The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.
[2026-01-23T04:54:15.508872Z] WARNING - Connection schemes (type: azure_data_factory) shall not contain '_' according to RFC3986.

[cruseakshay]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

hook-level test that covers the context manager behavior: AzureDataFactoryAsyncHook has test_context_manager_calls_close that explicitly verifies async with hook: calls close(). Haven't seen any trigger tests explicitly assert __aenter__/__aexit__ calls.

[dabla]

@cruseakshay Thanks for the PR and your patience, good job!

[boring-cyborg]

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.

[cruseakshay]

Closes: #55724

Fix unclosed aiohttp.ClientSession and TCPConnector warnings when using AzureDataFactoryRunPipelineOperator with deferrable=True.

The AzureDataFactoryAsyncHook creates an AsyncDataFactoryManagementClient but never closes it, causing resource leak errors in logs regardless of pipeline success/failure.

Changes

• Add __aenter__, __aexit__, and close() methods to AzureDataFactoryAsyncHook
• Update triggers to use async with hook: for proper cleanup
• Add unit tests for context manager behavior

Follows the same pattern as #52119 (Databricks fix).

---

Was generative AI tooling used to co-author this PR?

• Yes (please specify the tool below)

Generated-by: Claude following the guidelines

---

[SameerMesiah97]

Looks good to me. Turning the async hook into a context manager and updating the trigger call sites fixes the aiohttp client leak cleanly. Tests look solid as well.

Not a blocker but I would suggest double checking if there any places outside this file where this hook is being called so that the leak does not persist after your fix.

[cruseakshay]

@SameerMesiah97 Double checked other places, looks good.

[dabla]

2 small nits but looking good!

[dabla]

Looking good to me.

[dabla]

@cruseakshay Did you actually tests this changes and can you confirm this works?

[dabla]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

[cruseakshay]

@cruseakshay Did you actually tests this changes and can you confirm this works?

Yes, here are trail logs from latest run

[2026-01-23T04:53:13.235944Z] INFO - Sleeping for 60. The pipeline state is Queued.
[2026-01-23T04:54:13.245821Z] INFO - Request URL: 'https://management.azure.com/subscriptions/REDACTED/resourceGroups/airflow-adf-test-rg/providers/Microsoft.DataFactory/factories/airflow-adf-test/pipelineruns/REDACTED?api-version=REDACTED'
Request method: 'GET'
Request headers:
'Accept': 'application/json'
'x-ms-client-request-id': '9001a594-f817-11f0-b40d-b64b850643a2'
'User-Agent': 'azsdk-python-mgmt-datafactory/9.2.0 Python/3.12.12 (Linux-6.12.54-linuxkit-aarch64-with-glibc2.36)'
'Authorization': 'REDACTED'
No body was attached to the request
[2026-01-23T04:54:13.973572Z] INFO - Response status: 200
Response headers:
'Cache-Control': 'no-cache'
'Pragma': 'no-cache'
'Content-Length': '816'
'Content-Type': 'application/json; charset=utf-8'
'Expires': '-1'
'x-ms-correlation-request-id': 'REDACTED'
'x-ms-operation-identifier': 'REDACTED'
'x-ms-ratelimit-remaining-subscription-reads': '249'
'x-ms-ratelimit-remaining-subscription-global-reads': 'REDACTED'
'x-ms-request-id': '59f3995c-910c-4abe-952a-5e3e8fa3bff3'
'x-ms-routing-request-id': 'REDACTED'
'Strict-Transport-Security': 'REDACTED'
'X-Content-Type-Options': 'REDACTED'
'X-Cache': 'REDACTED'
'X-MSEdge-Ref': 'Ref A: 1C93C26BC61B478A8D2F48098F9A042C Ref B: MAA201060513029 Ref C: 2026-01-23T04:54:13Z'
'Date': 'Fri, 23 Jan 2026 04:54:13 GMT'
[2026-01-23T04:54:13.974417Z] INFO - Trigger fired eventname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) result=TriggerEvent<{'status': 'success', 'message': 'The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.', 'run_id': '66bb30ba-f817-11f0-bec5-16db0266592f'}> [2026-01-23T04:54:13.975233Z] INFO - trigger completedname=test_adf_trigger_fix/manual__2026-01-23T04:53:02.479036+00:00/run_pipeline_deferrable/-1/1 (ID 1) [2026-01-23T04:54:15.380187Z] INFO - DAG bundles loaded: dags-folder
[2026-01-23T04:54:15.380487Z] INFO - Filling up the DagBag from /usr/local/airflow/dags/test_adf_trigger_fix.py
[2026-01-23T04:54:15.488568Z] INFO - TaskInstance Details: dag_id=test_adf_trigger_fix task_id=run_pipeline_deferrable dagrun_id=manual__2026-01-23T04:53:02.479036+00:00 map_index=-1 run_start_date=2026-01-23T04:54:15.366708Z try_number=1 op_classpath=["airflow.providers.microsoft.azure.operators.data_factory.AzureDataFactoryRunPipelineOperator"] [2026-01-23T04:54:15.488765Z] INFO - The pipeline run 66bb30ba-f817-11f0-bec5-16db0266592f has Succeeded.
[2026-01-23T04:54:15.508872Z] WARNING - Connection schemes (type: azure_data_factory) shall not contain '_' according to RFC3986.

[cruseakshay]

@cruseakshay Also, the changes made in the data_factory trigger, like the added _build_trigger_event method like I asked, would be nice to see a test for it as well. Is the trigger data_factory tests also testing the context manger?

hook-level test that covers the context manager behavior: AzureDataFactoryAsyncHook has test_context_manager_calls_close that explicitly verifies async with hook: calls close(). Haven't seen any trigger tests explicitly assert __aenter__/__aexit__ calls.

[dabla]

@cruseakshay Thanks for the PR and your patience, good job!

[boring-cyborg]

Awesome work, congrats on your first merged pull request! You are invited to check our Issue Tracker for additional contributions.