Add Phoenix and TruLens third-party scorer integrations by debu-sinha · Pull Request #19237 · mlflow/mlflow

debu-sinha · 2025-12-05T07:08:28Z

Related Issues/PRs

Partial fix for #19062 (TruLens portion)

Note: This PR has been split per reviewer feedback. Phoenix scorers are now in #19473.

What changes are proposed in this pull request?

This PR adds integration for TruLens evaluation framework as MLflow GenAI scorers.

TruLens scorers (mlflow.genai.scorers.trulens):

Groundedness: Evaluates groundedness in context
ContextRelevance: Assesses context relevance to query
AnswerRelevance: Evaluates answer relevance to query
Coherence: Evaluates coherence and logical flow

Implementation details:

Lazy loading to avoid import overhead when scorers not used
Returns CategoricalRating.YES/NO values with scores in metadata
get_scorer() API for dynamic metric selection
Configurable model providers (OpenAI, Databricks, LiteLLM)
Trace input support for extracting context from retrieval spans
Clear error messages when optional dependencies are missing

Usage Examples with Real Output

Direct Scorer Call

from mlflow.genai.scorers.trulens import Groundedness

scorer = Groundedness(model="openai:/gpt-4o-mini")
feedback = scorer(
    outputs="The Eiffel Tower is 330 meters tall.",
    expectations={"context": "The Eiffel Tower is 330 metres tall and located in Paris."},
)
print(f"Value: {feedback.value}")
print(f"Metadata: {feedback.metadata}")

Output:

Value: yes
Metadata: {'mlflow.scorer.framework': 'trulens', 'score': 1.0, 'threshold': 0.5}

With mlflow.genai.evaluate()

import mlflow
from mlflow.genai.scorers.trulens import Groundedness, AnswerRelevance

eval_data = [
    {
        "inputs": {"query": "What is the capital of France?"},
        "outputs": "Paris is the capital of France.",
        "expectations": {"context": "France is in Western Europe. Its capital is Paris."},
    },
    {
        "inputs": {"query": "What is the capital of France?"},
        "outputs": "London is the capital of France.",  # Wrong - should fail
        "expectations": {"context": "France is in Western Europe. Its capital is Paris."},
    },
]

scorers = [
    Groundedness(model="openai:/gpt-4o-mini"),
    AnswerRelevance(model="openai:/gpt-4o-mini"),
]

results = mlflow.genai.evaluate(data=eval_data, scorers=scorers)
print(results.metrics)

Output:

{'Groundedness/mean': 0.5, 'AnswerRelevance/mean': 0.5}

All 4 TruLens Scorers

from mlflow.genai.scorers.trulens import Groundedness, ContextRelevance, AnswerRelevance, Coherence

all_scorers = [
    Groundedness(model="openai:/gpt-4o-mini"),
    ContextRelevance(model="openai:/gpt-4o-mini"),
    AnswerRelevance(model="openai:/gpt-4o-mini"),
    Coherence(model="openai:/gpt-4o-mini"),
]
results = mlflow.genai.evaluate(data=eval_data, scorers=all_scorers)

Output:

AnswerRelevance/mean: 0.5000
Coherence/mean: 0.5000
ContextRelevance/mean: 1.0000
Groundedness/mean: 0.5000

get_scorer() API

from mlflow.genai.scorers.trulens import get_scorer

groundedness = get_scorer("Groundedness", model="openai:/gpt-4o-mini")
context_relevance = get_scorer("ContextRelevance", model="databricks")
answer_relevance = get_scorer("AnswerRelevance", model="databricks:/my-endpoint")
coherence = get_scorer("Coherence", model="openai:/gpt-4o-mini")

How is this PR tested?

New unit tests in tests/genai/scorers/trulens/test_trulens.py
Real API integration tests with OpenAI (all tests passed)

Test coverage includes:

All 4 scorers with positive and negative test cases
Edge cases (empty strings, long text, special characters)
get_scorer() API for all metrics
mlflow.genai.evaluate() batch evaluation
CategoricalRating.YES/NO value validation
AssessmentSource.LLM_JUDGE type validation
Error handling for missing required fields

Does this PR require documentation update?

Yes. I've updated:
- API references with comprehensive docstrings

Release Notes

Is this a user-facing change?

Yes

Added TruLens third-party scorer integration for MLflow GenAI evaluation. 4 new scorers available:

TruLens scorers: Groundedness, ContextRelevance, AnswerRelevance, Coherence

Install dependencies:

pip install trulens trulens-providers-openai

What component(s), interfaces, languages, and integrations does this PR affect?

Components

area/evaluation: MLflow model evaluation features

How should the PR be classified in the release notes?

rn/feature - A new user-facing feature worth mentioning in the release notes

Should this PR be included in the next patch release?

Yes
No (this PR will be included in the next minor release)

github-actions · 2025-12-05T07:08:39Z

@debu-sinha Thank you for the contribution! Could you fix the following issue(s)?

⚠ Invalid PR template

This PR does not appear to have been filed using the MLflow PR template. Please copy the PR template from here and fill it out.

debu-sinha · 2025-12-06T03:46:42Z

@AveshCSingh Would you be able to take a look at this PR when you have a chance? It adds third-party scorer integrations for Phoenix (Arize) and TruLens evaluation frameworks, enabling seamless use of established LLM evaluation tools within MLflow's GenAI evaluation pipeline. Happy to address any feedback.

joelrobin18 · 2025-12-06T15:03:41Z

mlflow/genai/scorers/__init__.py

 }

+# Third-party scorer integrations (Phoenix, TruLens)
+_THIRDPARTY_IMPORTS = {


Instead of third party, it would be better to split them into two different folders
One for Phoenix and other for Trulens

That would be better instead of using third party.

Thanks for the suggestion @joelrobin18! I've restructured the code as you recommended - Phoenix and TruLens now have their own separate folders:

mlflow/genai/scorers/phoenix/

mlflow/genai/scorers/trulens/

Each integration has its own __init__.py with proper exports. Updated the lazy loading in the main scorers/__init__.py accordingly.

Let me know if there's anything else you'd like me to adjust!

debu-sinha · 2025-12-07T17:31:33Z

Hi @joelrobin18,

Thanks for the feedback on the folder structure. I've restructured the code as suggested:

Created separate mlflow/genai/scorers/phoenix/ directory
Created separate mlflow/genai/scorers/trulens/ directory
Each integration has its own __init__.py with proper exports
Updated the lazy loading in the main scorers/__init__.py accordingly

Would appreciate your re-review when you get a chance. Also cc @AveshCSingh for visibility.

This PR now focuses exclusively on TruLens agent trace scorers for goal-plan-action alignment evaluation. Basic TruLens scorers (Groundedness, ContextRelevance, AnswerRelevance, Coherence) are already provided in PR mlflow#19237 (Phoenix/TruLens third-party scorer integrations). Changes: - Remove mlflow/genai/scorers/trulens/basic.py (moved to mlflow#19237) - Update trulens/__init__.py with comprehensive examples for all 6 agent trace scorers - Update scorers/__init__.py to only export agent trace scorers - Update tests to only test agent trace scorers (19 tests remain) Agent trace scorers provided: - TruLensLogicalConsistencyScorer - TruLensExecutionEfficiencyScorer - TruLensPlanAdherenceScorer - TruLensPlanQualityScorer - TruLensToolSelectionScorer - TruLensToolCallingScorer 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> Signed-off-by: debu-sinha <debusinha2009@gmail.com>

This PR now focuses exclusively on TruLens agent trace scorers for goal-plan-action alignment evaluation. Basic TruLens scorers (Groundedness, ContextRelevance, AnswerRelevance, Coherence) are already provided in PR mlflow#19237 (Phoenix/TruLens third-party scorer integrations). Changes: - Remove mlflow/genai/scorers/trulens/basic.py (moved to mlflow#19237) - Update trulens/__init__.py with comprehensive examples for all 6 agent trace scorers - Update scorers/__init__.py to only export agent trace scorers - Update tests to only test agent trace scorers (19 tests remain) Agent trace scorers provided: - TruLensLogicalConsistencyScorer - TruLensExecutionEfficiencyScorer - TruLensPlanAdherenceScorer - TruLensPlanQualityScorer - TruLensToolSelectionScorer - TruLensToolCallingScorer Signed-off-by: debu-sinha <debusinha2009@gmail.com>

debu-sinha · 2025-12-11T20:12:20Z

Note: I also contributed a fix to the TruLens project (truera/trulens#2308) to add additionalProperties: false to Pydantic schemas for Databricks structured output compatibility. This was merged today and ensures TruLens works correctly with Databricks endpoints - relevant context for this integration.

debu-sinha · 2025-12-12T18:10:37Z

Update: Based on feedback from TruLens maintainers on the integration issue I opened (truera/trulens#2302), they requested adding their goal-plan-action alignment evaluators for agent trace analysis. I've created a follow-up PR for this: #19328

The TruLens team has been collaborative throughout - they merged my companion PR (truera/trulens#2308) to fix Databricks compatibility, and @sfc-gh-nvytla approved the integration on #19328.

B-Step62 · 2025-12-16T05:17:20Z

mlflow/genai/scorers/phoenix/__init__.py

@@ -0,0 +1,51 @@
+"""


The docstring in the module will be shown in API doc, but not visible from the doc website. I think we can keep this simple like #19345 and file a follow-up PR to have a proper documentation page. @smoorjani is adding doc for DeepEval scorers in #19409, so let's address this once the PR is merged.

My current docs PR is here: #19409
I'll be adding another one for RAGAS, but yes we can do this in a follow-up.

B-Step62 · 2025-12-16T05:19:11Z

mlflow/genai/scorers/phoenix/phoenix.py

+        from phoenix.evals import OpenAIModel
+
+        return OpenAIModel(model=self.model_name or "gpt-4o-mini")
+


Phoenix support several different models not only OpenAI: https://arize-phoenix.readthedocs.io/en/arize-phoenix-v4.10.1/api/evals.models.html

Can we support them? LiteLLM model will be fallback for models that are not natively supported by them e.g., Gemini.

B-Step62 · 2025-12-16T05:19:41Z

mlflow/genai/scorers/phoenix/phoenix.py

+        """Get the Phoenix OpenAI model instance."""
+        _check_phoenix_installed()
+
+        from phoenix.evals import OpenAIModel


We need to support Databricks judge and model serving endpoint. Could you read #19345 and apply the same approach?

B-Step62 · 2025-12-16T05:22:43Z

mlflow/genai/scorers/phoenix/phoenix.py

+                )
+                normalized_score = min(1.0, max(0.0, normalized_score))
+        else:
+            normalized_score = 1.0 if label == positive_label else 0.0


Can we use CategoricalRating.YES and CategoricalRating.NO?

B-Step62 · 2025-12-16T05:23:38Z

mlflow/genai/scorers/phoenix/phoenix.py

+            if normalized_score < 0.0 or normalized_score > 1.0:
+                import logging
+
+                logging.getLogger(__name__).warning(


Can we define logger at top level so we can reuse?

import logging logger = logging.getLogger(__name__) class _PhoenixScorerBase(Scorer): ...

B-Step62 · 2025-12-16T05:24:21Z

mlflow/genai/scorers/phoenix/phoenix.py

+
+                logging.getLogger(__name__).warning(
+                    f"Phoenix returned score {normalized_score} outside expected 0-1 range. "
+                    "This may indicate a version incompatibility. Clamping to valid range."


Do we need to crop the score to 0-1?

Good question - I've added clamping with a warning log. Phoenix metrics typically return 0-1, but defensive clamping ensures we don't pass unexpected values downstream. The warning helps surface any edge cases during debugging.

Do we have any example on when the score wasn't between 0-1?

B-Step62 · 2025-12-16T05:25:47Z

mlflow/genai/scorers/phoenix/phoenix.py

+        evaluator = HallucinationEvaluator(model=model)
+
+        # Build record dict as expected by Phoenix
+        query = inputs.get("query", str(inputs)) if inputs else ""


We have to handle more input patterns (e.g. trace, messages, etc). Can you update the logic similarly to the map_scorer_inputs_to_ragas_sample in #18988?

B-Step62 · 2025-12-16T05:26:40Z

mlflow/genai/scorers/phoenix/phoenix.py

+        context: str | None = None,
+        **kwargs,


A scorer should only have inputs, outputs, expectations, and trace as arguments. Other things like context should be derived from one of these.

B-Step62 · 2025-12-16T05:26:53Z

mlflow/genai/scorers/phoenix/phoenix.py

+        # Already aligned with MLflow convention (higher = better)
+        score, rationale = self._parse_result(result, positive_label="factual")
+
+        return Feedback(name=self.name, value=score, rationale=rationale)


Let's set assessment source as well.

B-Step62 · 2025-12-16T05:28:09Z

mlflow/genai/scorers/phoenix/phoenix.py

Can we also support get_judge() API?

small update - we called it get_scorer() now

smoorjani

left some nits in addition to Yuki's comments - I think the general ask here is to follow the existing pattern from DeepEval/RAGAS where we have a generalized implementation for any scorer (get_scorer) and the classes themselves (e.g., FaithfulnessScorer) are super lightweight wrappers.

smoorjani · 2025-12-16T22:18:16Z

mlflow/genai/scorers/__init__.py

        get_all_scorers,
    )
+    from mlflow.genai.scorers.phoenix import (
+        PhoenixHallucinationScorer,


nit: can we follow the existing pattern of mlflow.genai.scorers.phoenix import HallucinationScorer so as to not repeat Phoenix. We can follow-up later if this is verbose and renamespace as mlflow.geani.scorers import PhoenixHallucinationScorer.

smoorjani · 2025-12-16T22:18:55Z

mlflow/genai/scorers/phoenix/__init__.py

@@ -0,0 +1,51 @@
+"""


My current docs PR is here: #19409
I'll be adding another one for RAGAS, but yes we can do this in a follow-up.

smoorjani · 2025-12-16T22:19:43Z

mlflow/genai/scorers/phoenix/phoenix.py

small update - we called it get_scorer() now

smoorjani · 2025-12-16T22:22:53Z

mlflow/genai/scorers/phoenix/phoenix.py

+    try:
+        import phoenix.evals  # noqa: F401
+
+        return True


nit: does this need to return anything?

This was for _check_phoenix_installed() - it either raises an exception or returns implicitly. In the refactored code, the check happens in models.py during model creation, so it fails fast with a clear error message before any evaluation runs.

smoorjani · 2025-12-16T22:23:10Z

mlflow/genai/scorers/phoenix/phoenix.py

+
+
+def _check_phoenix_installed():
+    """Check if phoenix.evals is installed and raise a helpful error if not."""


nit: let's avoid one-line docstring, I think the linter will complain about this anyhow

smoorjani · 2025-12-16T22:25:50Z

mlflow/genai/scorers/phoenix/phoenix.py

+
+        return OpenAIModel(model=self.model_name or "gpt-4o-mini")
+
+    def _parse_result(


maybe this should be _construct_mlflow_assessment and this should return an assessment?

Agreed - I've refactored this entirely. The method now directly constructs and returns a object with proper . No separate helper method needed since each scorer's evaluation logic is self-contained.

debu-sinha · 2025-12-17T04:45:55Z

Thanks for the detailed feedback @B-Step62 @smoorjani! I've addressed all the comments in the latest push:

Structural changes:

Added get_scorer() API following the RAGAS pattern
Created lightweight wrapper classes: Hallucination, Relevance, Toxicity, QA, Summarization (without prefix)
Separated concerns into models.py, registry.py, utils.py for each integration

Model support:

Databricks managed judge: model="databricks"
Databricks serving endpoints: model="databricks:/<endpoint>"
Native providers: OpenAI, Azure, Bedrock, Anthropic, Gemini, Mistral
LiteLLM fallback for unsupported providers

Interface standardization:

Scorer args: inputs, outputs, expectations, trace only
Context derived from expectations or trace retrieval spans
CategoricalRating.YES/NO for binary feedback
Proper AssessmentSource attribution
Score clamping to 0-1 with warning logs

Ready for another look!

debu-sinha · 2025-12-17T19:29:13Z

Renewed tests on Databricks:

Implements support for Phoenix (Arize) and TruLens evaluation frameworks as MLflow GenAI scorers, enabling seamless integration of established LLM evaluation tools within MLflow's evaluation pipeline. Phoenix scorers (mlflow/genai/scorers/phoenix/): - PhoenixHallucinationScorer: Detects hallucinations in model outputs - PhoenixRelevanceScorer: Evaluates response relevance to queries - PhoenixToxicityScorer: Assesses content toxicity - PhoenixQAScorer: Evaluates QA correctness - PhoenixSummarizationScorer: Assesses summarization quality TruLens scorers (mlflow/genai/scorers/trulens/): - TruLensGroundednessScorer: Evaluates groundedness in context - TruLensContextRelevanceScorer: Assesses context relevance - TruLensAnswerRelevanceScorer: Evaluates answer relevance - TruLensCoherenceScorer: Evaluates logical flow of outputs Features: - Lazy loading to avoid import overhead when not used - Configurable model providers (OpenAI, LiteLLM for TruLens) - Consistent scorer interface returning Feedback objects - Helpful error messages when optional dependencies missing Signed-off-by: debu-sinha <debusinha2009@gmail.com>

Key changes: 1. Phoenix scorers: Remove incorrect score inversion for Hallucination and Toxicity evaluators. Phoenix natively returns scores aligned with MLflow convention (1.0 = good, 0.0 = bad). 2. Both Phoenix and TruLens scorers: Replace silent score clamping with validation that logs warnings when scores are outside expected 0-1 range. This helps detect potential version incompatibilities. 3. Update tests to reflect correct Phoenix score semantics: - factual = 1.0 (not inverted from 0) - non-toxic = 1.0 (not inverted from 0) 4. Update docstrings and comments to accurately describe score semantics. Signed-off-by: debu-sinha <debusinha2009@gmail.com>

- Align with existing DeepEval/RAGAS patterns for consistency - Add get_scorer() API for both Phoenix and TruLens integrations - Create lightweight wrapper classes (Hallucination, Relevance, etc.) - Add multi-model support: Databricks managed judge, serving endpoints, LiteLLM - Use standard scorer arguments: inputs, outputs, expectations, trace - Add AssessmentSource for proper attribution - Use CategoricalRating.YES/NO for binary feedback values - Add models.py, registry.py, utils.py for each integration - Update tests to match new API patterns Signed-off-by: debu-sinha <debusinha2009@gmail.com>

- Use function-level patching instead of module-level mocking - Import classes before applying mocks to ensure module is loaded - Fix line length issues in test files Signed-off-by: debu-sinha <debusinha2009@gmail.com>

Phoenix ToxicityEvaluator expects text in the 'input' field, but MLflow scorers typically pass text via 'outputs' parameter. When Toxicity is called with only outputs (no inputs), the text should be mapped to Phoenix's 'input' field. This fix ensures Toxicity()(outputs='text') works correctly by mapping the output to Phoenix's expected 'input' field when no input is provided. Signed-off-by: debu-sinha <debusinha2009@gmail.com>

The call_chat_completions function from databricks.rag_eval requires the @context.eval_context decorator which sets up internal state. When called directly from Phoenix/TruLens model adapters without this context, it fails with 'cannot access local variable' errors. Switch to using _invoke_databricks_serving_endpoint with the databricks-meta-llama-3-3-70b-instruct foundation model endpoint, which works in both notebook and external environments. Signed-off-by: debu-sinha <debusinha2009@gmail.com>

Phoenix evaluators use a set_verbosity context manager that expects model adapters to have both _verbose and _rate_limiter._verbose attributes. Without these, the context manager fails with UnboundLocalError. Added: - _NoOpRateLimiter stub class with _verbose attribute - _verbose = False to both Databricks model adapter classes - _rate_limiter = _NoOpRateLimiter() to both classes Signed-off-by: debu-sinha <debusinha2009@gmail.com>

Phoenix: - Add _verbose attribute required by set_verbosity context manager - Add _rate_limiter with _NoOpRateLimiter stub for set_verbosity TruLens: - Create dynamic provider class that inherits from LLMProvider - Implement _create_chat_completion method required by TruLens - TruLens feedback methods (groundedness, relevance, etc.) are inherited from LLMProvider base class Signed-off-by: debu-sinha <debusinha2009@gmail.com>

Phoenix fix: - Handle MultimodalPrompt objects by converting to string before passing to Databricks endpoint (Phoenix evaluators pass MultimodalPrompt, not str) TruLens fix: - Properly initialize Pydantic base class with super().__init__(model_engine=...) - Use class-level attributes for endpoint config to avoid Pydantic field issues Signed-off-by: debu-sinha <debusinha2009@gmail.com>

Add Endpoint object to TruLens Databricks provider initialization. TruLens LLMProvider requires an endpoint to be set for feedback methods like groundedness_measure_with_cot_reasons to work correctly. Signed-off-by: debu-sinha <debusinha2009@gmail.com>

debu-sinha · 2025-12-17T21:58:54Z

Hi @B-Step62,

Thanks for the feedback! I've split this PR as requested:

Phoenix PR: Add Phoenix (Arize) third-party scorer integration #19473 (ready for review)
TruLens PR: Will be created after Phoenix merges

The Phoenix PR includes:

All 5 Phoenix scorers (Hallucination, Relevance, Toxicity, QA, Summarization)
Simplified module docstrings per your feedback
get_scorer() API following RAGAS pattern
Trace input support using trace_utils
Databricks model support (managed judge and serving endpoints)
CategoricalRating.YES/NO values
AssessmentSource with LLM_JUDGE type

Please review #19473 when you have a chance. I'll close this PR and create a separate TruLens PR once Phoenix is merged.

github-actions bot added area/evaluation MLflow Evaluation rn/feature Mention under Features in Changelogs. labels Dec 5, 2025

debu-sinha force-pushed the feature/phoenix-trulens-judge-integrations branch 3 times, most recently from 09faf86 to 9a740c6 Compare December 6, 2025 15:02

joelrobin18 reviewed Dec 6, 2025

View reviewed changes

debu-sinha force-pushed the feature/phoenix-trulens-judge-integrations branch from 9a740c6 to 2e0cb24 Compare December 6, 2025 15:13

debu-sinha requested a review from joelrobin18 December 7, 2025 16:39

This was referenced Dec 11, 2025

[FR] Add TruLens Agent Trace Evaluators (Goal-Plan-Action Alignment) #19327

Open

Add TruLens agent trace scorers for goal-plan-action alignment evaluation #19328

Closed

debu-sinha force-pushed the feature/phoenix-trulens-judge-integrations branch from 255af44 to a0fff3b Compare December 11, 2025 09:29

B-Step62 added the v3.8.0 label Dec 12, 2025

B-Step62 requested review from AveshCSingh and smoorjani December 12, 2025 02:09

debu-sinha closed this Dec 12, 2025

debu-sinha reopened this Dec 12, 2025

debu-sinha mentioned this pull request Dec 12, 2025

[FR] Langchain trace cached tokens #18711

Open

14 tasks

B-Step62 reviewed Dec 16, 2025

View reviewed changes

github-actions bot assigned B-Step62 Dec 16, 2025

smoorjani requested changes Dec 16, 2025

View reviewed changes

github-actions bot assigned smoorjani Dec 17, 2025

debu-sinha force-pushed the feature/phoenix-trulens-judge-integrations branch from a33545e to 4938e31 Compare December 17, 2025 18:35

debu-sinha added 10 commits December 17, 2025 16:47

Fix Phoenix and TruLens test mocking

34bf17a

- Use function-level patching instead of module-level mocking - Import classes before applying mocks to ensure module is loaded - Fix line length issues in test files Signed-off-by: debu-sinha <debusinha2009@gmail.com>

debu-sinha force-pushed the feature/phoenix-trulens-judge-integrations branch from 4b1eb0a to a1058df Compare December 17, 2025 21:47

debu-sinha requested review from B-Step62 and smoorjani December 17, 2025 21:53

debu-sinha mentioned this pull request Dec 17, 2025

Add Phoenix (Arize) third-party scorer integration #19473

Merged

9 tasks

B-Step62 removed the v3.8.0 label Dec 18, 2025

B-Step62 closed this Dec 18, 2025

		from phoenix.evals import OpenAIModel

		return OpenAIModel(model=self.model_name or "gpt-4o-mini")



		def _check_phoenix_installed():
		"""Check if phoenix.evals is installed and raise a helpful error if not."""


		return OpenAIModel(model=self.model_name or "gpt-4o-mini")

		def _parse_result(

Conversation

debu-sinha commented Dec 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Related Issues/PRs

What changes are proposed in this pull request?

Usage Examples with Real Output

Direct Scorer Call

With mlflow.genai.evaluate()

All 4 TruLens Scorers

get_scorer() API

How is this PR tested?

Does this PR require documentation update?

Release Notes

Is this a user-facing change?

What component(s), interfaces, languages, and integrations does this PR affect?

How should the PR be classified in the release notes?

Should this PR be included in the next patch release?

Uh oh!

github-actions bot commented Dec 5, 2025

⚠ Invalid PR template

Uh oh!

debu-sinha commented Dec 6, 2025

Uh oh!

joelrobin18 Dec 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

debu-sinha commented Dec 7, 2025

Uh oh!

debu-sinha commented Dec 11, 2025

Uh oh!

debu-sinha commented Dec 12, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

smoorjani left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

debu-sinha commented Dec 5, 2025 •

edited

Loading

joelrobin18 Dec 6, 2025 •

edited

Loading

debu-sinha Dec 17, 2025 •

edited

Loading