Repository navigation
Description
This is RC2 for the three providers deferred from the October 20 release. Previous test confirmations are carried forward where applicable; new changes and the Azure durable-operator fixes need testing against RC2.
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 common.ai: 1.1.0rc2
- Publish usage, run id and model name from LLMOperator's subclasses (#75208): @pierrejeambrun
Linked issues: - common-ai: let a generic pydanticai connection set a default model provider (#75909): @jlaneve
Linked issues: - Add async aget_conn and acreate_agent to PydanticAIHook (#74150): @dabla
- Let Claude Agent SDK agent loops use common.ai toolsets through a new AirflowTools adapter (#74026): @Lee-W
- Support multimodal prompts with require_approval in LLM operators (#74388): @guan404ming
- Add emit_message_history to record an agent transcript without seeding it (#74811): @pankajastro
- Push LLMOperator's token usage and cost to XCom (#74355): @pierrejeambrun
- Make common.ai hooks report misconfiguration instead of failing silently (#74451): @Lee-W
- Load Common AI plugin tabs when the browser reaches Airflow on another origin (#75522): @kaxil
Linked issues: - Fix Common AI's AI Model tab for tasks inside a task loop (#75518): @kaxil
Linked issues: - Fix pulling AgentOperator output for list, marker and function types (#74729): @FrankYang0529
Linked issues: - Use
mask_secretfromcommon.compat1.22.0 in mcp hook (#75064): @mattwang44 - [main] Upgrade important CI environment (#73822): @github-actions[bot]
Linked issues: - Keep common.ai on anthropic 1.x until pydantic-ai supports anthropic 2 (#75640): @Lee-W
- List Meta among common.ai's external services and recognise new pydantic-ai 2.59 providers (#75641): @Lee-W
- Make LoggingToolset private and update Experimental notes in common.ai (#74694): @Lee-W
- Rewrite the Common AI sandbox configuration page around what users set up (#75826): @kaxil
- Add an LLM task loop example that stops on a dollar or token budget (#75516): @kaxil
Provider microsoft.azure: 15.3.0rc2
- Add AzureAISearchHook to the Microsoft Azure provider (#74290): @dabla
- Make the PowerBI refresh and MSGraph operators durable (#74675): @dabla
- Fix the durable MSGraph and PowerBI operators from post-merge review (#75272): @dabla
Linked issues: - add WasbDagBundle to load Dags from Azure Blob Storage (#67016): @Nishieee
Linked issues: - Fix MSGraphSensor timeout being reset on every re-defer (#75774): @stegololz
- Detect when enhanced Power BI refreshes finish or fail (#74600): @omar-massfih
Linked issues: - [main] Upgrade important CI environment (#73822): @github-actions[bot]
Linked issues: - Add deadline callback log visibility to the Dag Run view in UI (#73580): @seanghaeli
Linked issues:
Provider openlineage: 2.21.0rc2
- Fix task loop information and tests for OpenLineage provider (#74752): @kacpermuda
Linked issues: - Propagate OpenLineage parent and root job types to triggered Dags (#75303): @kacpermuda
- Allow excluding datasets from OpenLineage events by pattern and hook (#71312): @ramitkataria
Linked issues: - Match file datasets in exclude_datasets without a doubled slash (#75921): @ramitkataria
Linked issues: - Send task inlets and outlets as JSON lists in OpenLineage facet (#74666): @frank-ye
Linked issues: - Snowflake, Databricks: emit query lineage via OpenLineage public API (#73655): @kacpermuda
- Attribute log, asset-event and asset-state writes to the task attempt (#74347): @ashb
- Fix OpenLineage task events lost for Dags with mapped tasks (#75860): @kacpermuda
Linked issues: - Use task instance end date as OpenLineage terminal event time (#75328): @kacpermuda
Linked issues: - Surface errors from OpenLineage database info lookups without losing lineage (#74805): @kacpermuda
All users involved in the PRs:
@frank-ye @kaxil @pankajastro @ramitkataria @FrankYang0529 @pierrejeambrun @Nishieee @guan404ming @mattwang44 @github-actions[bot] @stegololz @ashb @Lee-W @seanghaeli @omar-massfih @dabla @kacpermuda @jlaneve
Drafted-by: Codex (GPT-6) (no human review before posting)
Activity
kaxil commentedon Oct 29, 2026
Common AI ones work
kacpermuda commentedon Oct 29, 2026
OL looks good as well, run e2e tests, some manual checks, nothing screams, loops work, thanks for rc2 !
hussein-awala commentedon Nov 4, 2026
Thank you to everyone who tested the release candidates and shared feedback!
Closing this testing issue now that the release is complete. Thanks again for helping validate it!