NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI, #486 most downloaded on PyPI
Datadog APM client library
Latest stable
4.17.2
Last release 3 days ago
19 Nov 2026
Ships on a steady schedule
a new release about every 2 weeks
Nearly every release is documented
notes for 60 of the last 60 stable releases
5 versions withdrawn
withdrawn after publishing
10 years old
871 releases · first in 2017
Drag across the ruler to compare, or tap a release to open it.
Moves ddtrace.provider.BaseContextProvider to ddtrace.trace.BaseContextProvider. The ddtrace.provider module is deprecated and will be removed in v3.0…
openai: Introduces tracing support to the OpenAI integration for Python versions 3.12 and 3.13.ddtrace.provider.BaseContextProvider to ddtrace.trace.BaseContextProvider. The ddtrace.provider module is deprecated and will be removed in v3.0.0.ddtrace.constants module:
ANALYTICS_SAMPLE_RATE_KEYSAMPLE_RATE_METRIC_KEYSAMPLING_PRIORITY_KEYSAMPLING_AGENT_DECISIONSAMPLING_RULE_DECISIONSAMPLING_LIMIT_DECISIONORIGIN_KEYUSER_ID_KEYHOSTNAME_KEYRUNTIME_FAMILYBASE_SERVICE_KEYSPAN_MEASURED_KEYKEEP_SPANS_RATE_KEYMULTIPLE_IP_HEADERSCONFIG_ENDPOINT_ENVCONFIG_ENDPOINT_RETRIES_ENVCONFIG_ENDPOINT_TIMEOUT_ENVddtrace.settings.config module and deprecates the following ddtrace.config attributes:
http, use DD_TRACE_HEADER_TAGS environment variable instead.http_server, use DD_TRACE_HTTP_SERVER_ERROR_STATUSES environment variable instead.trace_headers, this attribute is internal to the tracer.header_is_traced, this attribute is internal to the tracer.convert_rc_trace_sampling_rules, this attribute is internal to the tracer.enable_remote_configuration, use DD_REMOTE_CONFIGURATION_ENABLED environment variable instead.get_from, use ddtrace.trace.Pin to set instance level configurations.CI Visibility
Code security
LLM Observability
LLMObs.enable() to configure LLM Observability without environment variables would not automatically propagate distributed tracing headers.openai: Fixes a patching issue where asynchronous moderation endpoint calls resulted in coroutine scheduling errors.vertexai: Resolves an issue with chat.send_message() where the content keyword argument was not parsed correctly.Profiling
cassandra,cherrypy,flask_cache,starlette: Ensures a deprecation warning is not raised when patching these integrations via ddtrace-run and import ddtrace.auto.Your coding agent can read these notes before it upgrades. Set up the MCP server →
Ensures that no module from ASM are loaded when ASM is disabled or unavailable.
ASM
ImportError exception on Windows.Code Security
Tracing
Moves ddtrace.provider.BaseContextProvider to ddtrace.trace.BaseContextProvider. The ddtrace.provider module is deprecated and will be removed in v3.0…
ddtrace.provider.BaseContextProvider to ddtrace.trace.BaseContextProvider. The ddtrace.provider module is deprecated and will be removed in v3.0.0.ddtrace.constants module:
ANALYTICS_SAMPLE_RATE_KEYSAMPLE_RATE_METRIC_KEYSAMPLING_PRIORITY_KEYSAMPLING_AGENT_DECISIONSAMPLING_RULE_DECISIONSAMPLING_LIMIT_DECISIONORIGIN_KEYUSER_ID_KEYHOSTNAME_KEYRUNTIME_FAMILYBASE_SERVICE_KEYSPAN_MEASURED_KEYKEEP_SPANS_RATE_KEYMULTIPLE_IP_HEADERSCONFIG_ENDPOINT_ENVCONFIG_ENDPOINT_RETRIES_ENVCONFIG_ENDPOINT_TIMEOUT_ENVddtrace.settings.config module and deprecates the following ddtrace.config attributes:
http, use DD_TRACE_HEADER_TAGS environment variable instead.http_server, use DD_TRACE_HTTP_SERVER_ERROR_STATUSES environment variable instead.trace_headers, this attribute is internal to the tracer.header_is_traced, this attribute is internal to the tracer.convert_rc_trace_sampling_rules, this attribute is internal to the tracer.enable_remote_configuration, use DD_REMOTE_CONFIGURATION_ENABLED environment variable instead.get_from, use ddtrace.trace.Pin to set instance level configurations.LLMObs.enable() to configure LLM Observability without environment variables would not automatically propagate distributed tracing headers.vertexai: Resolves an issue with chat.send_message() where the content keyword argument was not parsed correctly.cassandra, cherrypy ,flask_cache, starlette: Ensures a deprecation warning is not raised when patching these integrations via ddtrace-run and import ddtrace.auto.Adds support for Header Injection vulnerability sink point.
ddtrace-run entrypoint.CI Visibility
ddtrace.contrib.<integration> to ddtrace.contrib.internal.<integration>.Tracing
ddtrace.filters.FilterRequestsOnUrl. Spans should be filtered/sampled using DD_TRACE_SAMPLING_RULES configuration.ddtrace.tracer) `should be used instead.ddtrace.configure(...) function. These parameters will be removed in ddtrace>=3.0.0: - enabled - hostname - port - uds_path - https - sampler - settings - priority_sampling - settings - dogstatsd_url - writer - partial_flush_enabled - partial_flush_min_spans - api_version - compute_stats_enabled - wrap_executorddtrace.pin module and moves the Pin class to ddtrace.trace package. In v3.0.0 the ddtrace/pin.py will be removed.ddtrace.filters module and moves the TraceFilter and FilterRequestsOnUrl classes to ddtrace.trace package. In v3.0.0 the ddtrace/filters.py will be removed.ddtrace.contrib.trace_utils_async and ddtrace.contrib.redis_utils. Replaces ddtrace.contrib.trace_utils_async.with_traced_module(...) with ddtrace.contrib.trace_utils.with_traced_module_async(...). Moves public attributes defined in ddtrace.contrib.redis_utils.* to ddtrace.contrib.trace_utils.ddtrace>=3.0.0 integrations should only be enabled and configured via ddtrace.patch(..), import ddtrace.auto or the ddtrace-run command. Unpatching integrations or getting the version of an integration is no longer supported.rq: Ensures the implementation details of the rq integration are internal to ddtrace library. In ddtrace>=3.0.0 this integration should only be enabled and configured via ddtrace.patch(..), import ddtrace.auto or the ddtrace-run commandASM
Code Security
LLM Observability
openai: Introduces automatic extraction of token usage from streamed chat completions. Unless stream_options: {"include_usage": False} is explicitly set on your streamed chat completion request, the OpenAI integration will add stream_options: {"include_usage": True} to your request and automatically extract the token usage chunk from the streamed response.Profiling
gunicorn with Stack V2 results in performance degradationgevent is lackingDD_PROFILING_STACK_V2_ENABLED=0. If you find any other issue, then please proceed to escalate using appropriate support channels or file an issue on the repository.Tracing
DD_TRACE_PROPAGATION_BEHAVIOR_EXTRACT to control the behavior of the extraction of distributed tracing headers. The values, continue (default), ignore, and restart, are supported. The default value is continue which has no change from the current behavior of always propagating valid headers. ignore ignores all incoming headers, never propagating the incoming trace information and restart turns the first extracted propagation style into a span link and propagates baggage if extracted.Code Security
Exception Replay
Lib Injection/SSI
LLM Observability
botocore: Resolves formatting errors in the bedrock integration when parsing request model IDs, which can now accept AWS ARNs.Profiling
AttributeError: _Profiled_ThreadingLock object has no attribute self_acquired_atTracing
asgi: Resolves an issue parsing response cookies in FastAPI and awsgiasyncio: Resolves an issue where asyncio event loops fail to register when ddtrace-run/import ddtrace.auto is used and gevent is installed.celery: Fixes an issue where celery.apply spans from Celery prerun got closed too soon leading to span tags being missing.django: Fixes issue where django cache is represented as a django service rather than the third party service.Adds support for Header Injection vulnerability sink point.
ddtrace-run entrypoint.CI Visibility
ddtrace.contrib.<integration> to ddtrace.contrib.internal.<integration>.Tracing
ddtrace.filters.FilterRequestsOnUrl. Spans should be filtered/sampled using DD_TRACE_SAMPLING_RULES configuration.ddtrace.tracer) `should be used instead.ddtrace.configure(...) function. These parameters will be removed in ddtrace>=3.0.0: - enabled - hostname - port - uds_path - https - sampler - settings - priority_sampling - settings - dogstatsd_url - writer - partial_flush_enabled - partial_flush_min_spans - api_version - compute_stats_enabled - wrap_executorddtrace.pin module and moves the Pin class to ddtrace.trace package. In v3.0.0 the ddtrace/pin.py will be removed.ddtrace.filters module and moves the TraceFilter and FilterRequestsOnUrl classes to ddtrace.trace package. In v3.0.0 the ddtrace/filters.py will be removed.ddtrace.contrib.trace_utils_async and ddtrace.contrib.redis_utils. Replaces ddtrace.contrib.trace_utils_async.with_traced_module(...) with ddtrace.contrib.trace_utils.with_traced_module_async(...). Moves public attributes defined in ddtrace.contrib.redis_utils.* to ddtrace.contrib.trace_utils.ddtrace>=3.0.0 integrations should only be enabled and configured via ddtrace.patch(..), import ddtrace.auto or the ddtrace-run command. Unpatching integrations or getting the version of an integration is no longer supported.rq: Ensures the implementation details of the rq integration are internal to ddtrace library. In ddtrace>=3.0.0 this integration should only be enabled and configured via ddtrace.patch(..), import ddtrace.auto or the ddtrace-run commandASM
Code Security
LLM Observability
openai: Introduces automatic extraction of token usage from streamed chat completions. Unless stream_options: {"include_usage": False} is explicitly set on your streamed chat completion request, the OpenAI integration will add stream_options: {"include_usage": True} to your request and automatically extract the token usage chunk from the streamed response.Profiling
gunicorn with Stack V2 results in performance degradationgevent is lackingDD_PROFILING_STACK_V2_ENABLED=0. If you find any other issue, then please proceed to escalate using appropriate support channels or file an issue on the repository.Tracing
DD_TRACE_PROPAGATION_BEHAVIOR_EXTRACT to control the behavior of the extraction of distributed tracing headers. The values, continue (default), ignore, and restart, are supported. The default value is continue which has no change from the current behavior of always propagating valid headers. ignore ignores all incoming headers, never propagating the incoming trace information and restart turns the first extracted propagation style into a span link and propagates baggage if extracted.Code Security
Exception Replay
Lib Injection/SSI
LLM Observability
botocore: Resolves formatting errors in the bedrock integration when parsing request model IDs, which can now accept AWS ARNs.Profiling
AttributeError: _Profiled_ThreadingLock object has no attribute self_acquired_atTracing
asgi: Resolves an issue parsing response cookies in FastAPI and awsgiasyncio: Resolves an issue where asyncio event loops fail to register when ddtrace-run/import ddtrace.auto is used and gevent is installed.celery: Fixes an issue where celery.apply spans from Celery prerun got closed too soon leading to span tags being missing.django: Fixes issue where django cache is represented as a django service rather than the third party service.ASM: This ensures that no module from ASM are loaded when ASM is disabled or unavailable.
ASM: This ensures that no module from ASM are loaded when ASM is disabled or unavailable.
SCA: This ensures that no module from IAST are loaded when IAST is disabled or unavailable.
Runtime Code Analysis (IAST): Avoid imports of IAST native module when IAST is not enabled.
ASM: This fix resolves an issue where IAST modules could be loaded, even if disabled,
which could create an ImportError exception on Windows.
CI Visibility: fixes an issue where Auto Test Retries with pytest would always consider retries of tests defined inside unittest classes to be success
celery: Fixes an issue where celery.apply spans from Celery prerun got closed too soon leading to span tags being missing.
celery.apply spans from Celery prerun got closed too soon leading to span tags being missing.chat.send_message() where the content keyword argument was not parsed correctly.Fixes an issue where the memory allocation profiler can cause a segmentation fault due to data races when accessing its own global data structures fro
Profiling
Tracing
django: Fixes issue where django cache is represented as a django service rather than the third party service.botocore: Resolves formatting errors in the bedrock integration when parsing request model IDs, which can now accept AWS ARNs.Introduces "Standalone SCA billing", opting out for APM billing and applying to only SCA. Enable this by setting these two environment variables: DD_A
ASM
DD_APPSEC_SCA_ENABLED and DD_EXPERIMENTAL_APPSEC_STANDALONE_ENABLEDCode Security
Profiling
DD_PROFILING_PYTORCH_ENABLED=true. This feature instruments the PyTorch profiler API (https://pytorch.org/docs/stable/_modules/torch/profiler/profiler.html) so that GPU profiling data can be sent to Datadog for visualization. This feature supports torch version >= 1.8.1.Tracing
azure_functions: Introduces support for Azure Functions.ASM
Code Security
Lib-Injection
LLM Observability
LLMObs.enable() ignored global patch configurations, specificallyDD_TRACE_<INTEGRATION>_ENABLED and DD_PATCH_MODULES environment variables.Telemetry
Tracing
ASGI: This fix resolves an issue parsing response cookies in FastAPI and awsgiasyncio: Resolves an issue where asyncio event loops fail to register when ddtrace-run/import ddtrace.auto is used and gevent is installed.datastreams: Logs at warning level for Kinesis errors that break the Data Streams Monitoring map.Adds more modules to the IAST patching denylist to improve startup time
Code Security
Tracing
ASGI: Resolves an issue parsing response cookies in FastAPI and awsgiasyncio: Resolves an issue where asyncio event loops fail to register when ddtrace-run/import ddtrace.auto is used and gevent is installed.Makes the library compatible with Python 3.13.
ASM
DD_APPSEC_SCA_ENABLED and DD_EXPERIMENTAL_APPSEC_STANDALONE_ENABLEDCode Security
Profiling
DD_PROFILING_PYTORCH_ENABLED=true. This feature instruments the PyTorch profiler API so that GPU profiling data can be sent to Datadog for visualization. This feature supports torch version >= 1.8.1.Tracing
azure_functions: Introduces support for Azure Functions.ASM
Lib-Injection
LLM Observability
LLMObs.enable() ignored global patch configurations, specificallyDD_TRACE_<INTEGRATION>_ENABLED and DD_PATCH_MODULES environment variables.Telemetry
Tracing
datastreams: Logs at warning level for Kinesis errors that break the Data Streams Monitoring map.Adds more modules to the IAST patching denylist to improve startup time
Code Security
Profiling
Tracing
ASGI: Resolves an issue parsing response cookies in FastAPI and awsgiFixes an issue where the memory allocation profiler can cause a segmentation fault due to data races when accessing its own global data structures fro
Profiling:
Tracing:
celery.apply spans from Celery prerun got closed too soon leading to span tags being missing.Ensures IAST SSRF vulnerability redacts the url query parameters correctly.
With this upgrade, you can now control how the stack trace report are cropped when reported for exploit prevention or IAST.
DD_APPSEC_MAX_STACK_TRACE_DEPTH allowed to control the maximum stack trace size reported (default 32)DD_APPSEC_MAX_STACK_TRACE_DEPTH_TOP_PERCENT allows now to specify how the stack trace is cropped as a percentage.For example, a value of 100 will report the top DD_APPSEC_MAX_STACK_TRACE_DEPTH frames from the stack, while a value of 0 will report the bottom DD_APPSEC_MAX_STACK_TRACE_DEPTH frames of the trace. A value of 50 will report half of DD_APPSEC_MAX_STACK_TRACE_DEPTH (rounded down) frames from the top of the stack and the rest from bottom. Default value is 75.
Upgrades libddwaf to 1.22.0
Upgrades libddwaf to 1.21.0 and security rule file to 1.13.3
CI Visibility
Beta release of the new version of the pytest plugin, introducing the following features:
Set the DD_PYTEST_USE_NEW_PLUGIN_BETA environment variable to true to use this new version.
NOTE: this new version of the plugin introduces breaking changes:
module, suite, and test names are now parsed from the item.nodeid attributeAdds support for Selenium and RUM integration
Code Security
DD_IAST_ENABLED=1 and DD_EXPERIMENTAL_APPSEC_STANDALONE_ENABLED=1.LLM Observability
vertexai: Introduces tracing support for Google's Vertex AI SDK for Python's generate_content and send_message calls. See the docs for more information.Profiling
tracer.configure()ASM
CI Visibility
None in EVP proxy mode if DD_ENV was not specified but the agent had a default environment set to a value other than none (eg: using DD_APM_ENV in the agent's environment).--ddtrace are no longer skipped when executing tests with pytest. Previously, the algorithm misinterpreted these arguments as standard flags, overlooking possible test paths that may contribute to the inferred service name.Code Security
psycopg3.umap, numba and pynndescent to the Code Security denylist.Crashtracking
Lib-Injection
sitecustomize.py are preserved when auto-injecting.click and jinja2 installed on 3.8 were outside of the allowed minimum versions for autoinstrumentation.LLM Observability
langchain: Resolves a JSON decoding issue resulting from tagging streamed outputs from chains ending with a PydanticOutputParser.Profiling
asyncio task names are not properly propagated when using stack v2, i.e. when DD_PROFILING_STACK_V2_ENABLED is set. Fixes an issue where asyncio tasks are not associated with spans when using stack v2, i.e. when DD_PROFILING_STACK_V2_ENABLED is set.Telemetry
Tracing
botocore: This fix resolves an issue in the Bedrock integration where not consuming the full response stream would prevent spans from finishing.botocore: This fix resolves the issue where the span pointer for deserialized DynamoDB requests (through the resource-based API) were not being generated.botocore: This fix resolves an issue where our span pointer calculation code added recently logged unactionable messages.celery: This fix resolves two issues with context propagation in celery
celery: Changes celery out.host span tag to point towards broker host url instead of local celery process hostname. Fixes inferred service representation issues when using celery.grpcaio: Resolves a concurrency bug where distributed tracing headers were overwritten resulting in spans being assigned to the wrong trace.kafka: Fixes an issue with Kafka consumer spans not using the active trace context when distributed tracing was enabled and no valid distributed context found was found within a consumed message.Ensures IAST SSRF vulnerability redacts the url query parameters correctly.
With this upgrade, you can now control how the stack trace report are cropped when reported for exploit prevention or IAST.
DD_APPSEC_MAX_STACK_TRACE_DEPTH allowed to control the maximum stack trace size reported (default 32)DD_APPSEC_MAX_STACK_TRACE_DEPTH_TOP_PERCENT allows now to specify how the stack trace is cropped as a percentage.For example, a value of 100 will report the top DD_APPSEC_MAX_STACK_TRACE_DEPTH frames from the stack, while a value of 0 will report the bottom DD_APPSEC_MAX_STACK_TRACE_DEPTH frames of the trace. A value of 50 will report half of DD_APPSEC_MAX_STACK_TRACE_DEPTH (rounded down) frames from the top of the stack and the rest from bottom. Default value is 75.
Upgrades libddwaf to 1.22.0
Upgrades libddwaf to 1.21.0 and security rule file to 1.13.3
CI Visibility
Beta release of the new version of the pytest plugin, introducing the following features:
Set the DD_PYTEST_USE_NEW_PLUGIN_BETA environment variable to true to use this new version.
NOTE: this new version of the plugin introduces breaking changes:
module, suite, and test names are now parsed from the item.nodeid attributeAdds support for Selenium and RUM integration
Code Security
-Introduces "Standalone Code Security", a feature that disables APM in the tracer but keeps Code Security (IAST) enabled. In order to enable it, set the environment variables DD_IAST_ENABLED=1 and DD_EXPERIMENTAL_APPSEC_STANDALONE_ENABLED=1.
LLM Observability
vertexai: Introduces tracing support for Google's Vertex AI SDK for Python's generate_content and send_message calls. See the docs for more information.Profiling
tracer.configure()ASM
CI Visibility
None in EVP proxy mode if DD_ENV was not specified but the agent had a default environment set to a value other than none (eg: using DD_APM_ENV in the agent's environment).--ddtrace are no longer skipped when executing tests with pytest. Previously, the algorithm misinterpreted these arguments as standard flags, overlooking possible test paths that may contribute to the inferred service name.Code Security
psycopg3.umap, numba and pynndescent to the Code Security denylist.Crashtracking
Lib-Injection
sitecustomize.py are preserved when auto-injecting.click and jinja2 installed on 3.8 were outside of the allowed minimum versions for autoinstrumentation.LLM Observability
langchain: Resolves a JSON decoding issue resulting from tagging streamed outputs from chains ending with a PydanticOutputParser.Profiling
asyncio task names are not properly propagated when using stack v2, i.e. when DD_PROFILING_STACK_V2_ENABLED is set. Fixes an issue where asyncio tasks are not associated with spans when using stack v2, i.e. when DD_PROFILING_STACK_V2_ENABLED is set.Telemetry
Ensures that Telemetry heartbeats are not skipped for forked processes, as doing so could result in the dependency list being lost over time.
Tracing
botocore: This fix resolves an issue in the Bedrock integration where not consuming the full response stream would prevent spans from finishing.botocore: This fix resolves the issue where the span pointer for deserialized DynamoDB requests (through the resource-based API) were not being generated.botocore: This fix resolves an issue where our span pointer calculation code added recently logged unactionable messages.celery: This fix resolves two issues with context propagation in celery
celery: Changes celery out.host span tag to point towards broker host url instead of local celery process hostname. Fixes inferred service representation issues when using celery.grpcaio: Resolves a concurrency bug where distributed tracing headers were overwritten resulting in spans being assigned to the wrong trace.kafka: Fixes an issue with Kafka consumer spans not using the active trace context when distributed tracing was enabled and no valid distributed context found was found within a consumed message.celery: Fixes an issue where celery.apply spans from Celery pre-run got closed too soon leading to span tags being missing.
celery: Fixes an issue where celery.apply spans from Celery pre-run got closed too soon leading to span tags being missing.Adds more modules to the IAST patching denylist to improve startup time
Code Security
ASM
LLM Observability
LLMObs.enable() ignored global patch configurations, specifically the DD_TRACE_<INTEGRATION>_ENABLED and DD_PATCH_MODULES environment variables.langchain: Resolves a JSON decoding issue resulting from tagging streamed outputs from chains ending with a PydanticOutputParser.Profiling
Tracing
ASGI: Resolves an issue parsing response cookies in FastAPI and awsgiEnsure IAST SSRF vulnerability redacts the url query parameters correctly.
SCA:
Celery:
Code Security:
Profiling:
Ensures that common patches for exploit prevention and SCA are only loaded if required, and only loaded once.
ASM
LLM Observability
Tracing
botocore: Resolves an issue in the Bedrock integration where not consuming the full response stream would prevent spans from finishing.celery: Changes celery out.host span tag to point towards broker host url instead of local celery process hostname. Fixes inferred service representation issues when using celery.grpcaio: Resolves a concurrency bug where distributed tracing headers were overwritten resulting in spans being assigned to the wrong trace.Resolves an issue where some root spans were not appropriately tagged for ASM standalone.
click and jinja2 installed on 3.8 were outside of the allowed minimum versions for autoinstrumentation.Support added for session fingerprints.
ASM
LLM Observability
llm span.chat_model.with_structured_output(..., method="json_mode") is used, or response_format={"type": "json_object"} is passed into a langchain chat model invocation, the LLM Observability span will be an llm span instead of a workflow span.SSI
requirements.json to SSI artifact for bailing out on unsupported systems.Tracing
DD_SERVICE is not set, replacing instances of 'unnamed-python-service'. Ensures that a more meaningful service name is used whenever possible, enhancing clarity in service identification.ASM
DD_DJANGO_INCLUDE_EMAIL (false by default), will tag user events with user email as before.Code Security/IAST
umap, numba and pynndescent to the Code Security denylist.googlecloudsdk and google auth to the Code Security deny list.google.cloud.storage.batch module would fail raising an ImportErrorCrashtracking
Lib-Injection
ddtrace if it is present.sys.argv is not available.LLM Observability
Profiling
DD_PROFILING_STACK_V2_ENABLED was set.DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_TIMELINE_ENABLED or DD_PROFILING_STACK_V2_ENABLED turned off live heap profiling.ensure_binary_or_empty() function, on Python versions before 3.12, with DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.fork() by clearing it in the child process.Span. The mapping is populated and used when DD_PROFILING_ENDPOINT_COLLECTION_ENABLED and DD_PROFILING_STACK_V2_ENABLED were set to enable grouping of profiles for endpoints.Tracing
--ddtrace are no longer skipped when executing tests with pytest. Previously, the algorithm misinterpreted these arguments as standard flags, overlooking possible test paths that may contribute to the inferred service name.botocore: Resolves the issue where the span pointer for deserialized DynamoDB requests (through the resource-based API) were not being generated.botocore: Resolves an issue where our span pointer calculation code added recently logged unactionable messages.pymongo: add type checking to solve an issue where NoneType instead of expected Pin object would throw an error in TracedTopology method.Adds umap, numba and pynndescent to the Code Security denylist.
Code Security
umap, numba and pynndescent to the Code Security denylist.Lib-Injection
Tracing
--ddtrace are no longer skipped when executing tests with pytest. Previously, the algorithm misinterpreted these arguments as standard flags, overlooking possible test paths that may contribute to the inferred service name.Adds support for session fingerprints.
ASM
LLM Observability
llm span.chat_model.with_structured_output(..., method="json_mode") is used, or response_format={"type": "json_object"} is passed into a langchain chat model invocation, the LLM Observability span will be an llm span instead of a workflow span.Single Step Instrumentation
requirements.json to SSI artifact for bailing out on unsupported systems.Tracing
DD_SERVICE is not set, replacing instances of unnamed-python-service. Ensures that a more meaningful service name is used whenever possible, enhancing clarity in service identification.ASM/Threats
DD_DJANGO_INCLUDE_EMAIL (false by default), will tag user events with user email as before.Code Security/IAST
google.cloud.storage.batch module would fail raising an ImportError.Crashtracking
Lib-Injection
ddtrace if it is present.sys.argv is not available.LLM Observability
Profiling
DD_PROFILING_STACK_V2_ENABLED was set.DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_TIMELINE_ENABLED or DD_PROFILING_STACK_V2_ENABLED turned off live heap profiling.ensure_binary_or_empty() function, on Python versions before 3.12, with DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.fork() by clearing it in the child process.Span. The mapping is populated and used when DD_PROFILING_ENDPOINT_COLLECTION_ENABLED and DD_PROFILING_STACK_V2_ENABLED were set to enable grouping of profiles for endpoints.Tracing
NoneType instead of expected Pin object would throw an error in TracedTopology method.Ensure IAST SSRF vulnerability redacts the url query parameters correctly.
SCA:
Code Security:
Lib-Injection:
Profiling
Ensures that common patches for exploit prevention and sca are only loaded if required, and only loaded once.
ASM
Auto-Instrumentation
click and jinja2 installed on python3.8 were outside of the allowed minimum versions for auto-instrumentation.Code Security
LLM Observability
Tracing
botocore: Resolves an issue in the Bedrock integration where not consuming the full response stream would prevent spans from finishing.botocore: Resolves the issue where the span pointer for deserialized DynamoDB requests (through the resource-based API) were not being generated.
Code Security: add umap, numba and pynndescent to the Code Security denylist.
The lock profiler would log a warning if it couldn't determine a name for a lock, and it would try determining a name multiple times for the same lock
Profiling
Tracing
NoneType instead of expected Pin object would throw an error in TracedTopology method.The new user events policy is preventing users PII to be added by default as span tags. To allow customers using the Django auto instrumentation to st
Threats
DD_DJANGO_INCLUDE_EMAIL (false by default), will tag user events with user email as before.Code Security
Crashtracking
Lib-Injection
ddtrace if it is present.LLM Observability
Profiling
DD_PROFILING_STACK_V2_ENABLED was set.ensure_binary_or_empty() function, on Python versions before 3.12, with DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.fork() by clearing it in the child process.When starting LLM and embedding spans, the model_name argument is now optional and will default to custom. This applies to both inline methods (e.g. L
LLM Observability
model_name argument is now optional and will default to custom. This applies to both inline methods (e.g. LLMObs.llm()) and function decorators (e.g. @llm).submit_evaluation method. For more information, see submitting evaluations with the SDK.Tracing
BatchWriteItem spans. Table Primary Keys will need to be provided with the ddtrace.config.botocore.dynamodb_primary_key_names_for_tables option or the DD_BOTOCORE_DYNAMODB_TABLE_PRIMARY_KEYS environment variable to correctly handle the PutRequest items.TransactWriteItems spans. Table Primary Keys will need to be provided with the ddtrace.config.botocore.dynamodb_primary_key_names_for_tables option or the DD_BOTOCORE_DYNAMODB_TABLE_PRIMARY_KEYS environment variable to correctly handle the Put items.ddtrace.config.botocore.add_span_pointers option or the DD_BOTOCORE_ADD_SPAN_POINTERS environment variable to control adding span pointers to some successful AWS API requests. This option is enabled by default.CI Visibility
CODEOWNERS would incorrectly fail to discard line-level trailing comments (eg: @code/owner # my comment would result in codeowners being parsed as @code/owner, #, my, and comment)git binary or .git directory)Code Security
google.cloud.storage.batch module would fail raising an ImportErrorDynamic Instrumentation
LLM Observability
LLMObs.annotate.Profiling
DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_TIMELINE_ENABLED or DD_PROFILING_STACK_V2_ENABLED turned off live heap profiling.Span. The mapping is populated and used when DD_PROFILING_ENDPOINT_COLLECTION_ENABLED and DD_PROFILING_STACK_V2_ENABLED were set to enable grouping of profiles for endpoints.DD_PROFILING_STACK_V2_ENABLED is set.TypeError exception when it is given a Span with None span_type.Tracing
AGENT_CONFIG remote configuration product.LLMObs.annotate is called multiple times on the same span so that the latest value for a tag key overrides the previous value.Nothing published for this version
Nothing published for this version
Ensures that common patches for exploit prevention and sca are only loaded if required, and only loaded once.
ASM
Auto-Instrumentation
click and jinja2 installed on python3.8 were outside of the allowed minimum versions for auto-instrumentation.Code Security
LLM Observability
Tracing
botocore: Resolves an issue in the Bedrock integration where not consuming the full response stream would prevent spans from finishing.The new user events policy is preventing users PII to be added by default as span tags. To allow customers using the Django auto instrumentation to st
ASM:
LLM Observability:
Code Security:
Profiling:
Fixes an issue where cpu-time was not profiled for services using gunicorn, when <span class="title-ref">`DD_PROFILING_STACK_V2_ENABLED</span> was set.
The lock profiler would log a warning if it couldn't determine a
name for a lock, and it would try determining a name multiple times for the same lock. This lead to excessive log spam. Downgrade this to a debug log and only try to determine the name once.
Fixes an issue where the sample pool could deadlock after fork()
by clearing it in the child process.
Fixes an issue where enabling native exporter via DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_TIMELINE_ENABLED or DD_PROFILING_STACK_V2_ENABLED tu
DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_TIMELINE_ENABLED or DD_PROFILING_STACK_V2_ENABLED turned off live heap profiling.ensure_binary_or_empty() function, on Python versions before 3.12, with DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.Span. The mapping is populated and used when DD_PROFILING_ENDPOINT_COLLECTION_ENABLED and DD_PROFILING_STACK_V2_ENABLED were set to enable grouping of profiles for endpoints.DD_PROFILING_STACK_V2_ENABLED is set.NoneType instead of expected Pin object would throw an error in TracedTopology method.Fixes a bug where CODEOWNERS would incorrectly fail to discard line-level trailing comments (eg: @code/owner # my comment would result in codeowners b
CODEOWNERS would incorrectly fail to discard line-level trailing comments (eg: @code/owner # my comment would result in codeowners being parsed as @code/owner, #, my, and comment)git binary or .git directory)google.cloud.storage.batch module would fail raising an ImportError.TypeError exception when it is given a Span with None span_type.Additionally, it improves validations in both the Processor and Vulnerability Reporter, depending on whether IAST is active or not.
LLM Observability
prompt and name arguments to LLMObs.annotation_context to support setting an integration generated span's name and prompt field. For more information on annotation contexts, see the docs here.stream calls on LCEL chains, chat completion models, or completion models. Note that due to an upstream issue with the langchain library itself, streamed responses will not be tagged correctly based on the choice index when the underlying model is configured to return n>1 choices. Please refer to this GitHub issue for more details.llm.stream, chat_model.stream, and chain.stream) submit to LLM Observability.CI Visibility
test_session.name tag to test events. The test session name can be set via the DD_TEST_SESSION_NAME environment variable. If DD_TEST_SESSION_NAME is not specified, the test session name is set from the CI job id and the test command.Tracing
DD_CODE_ORIGIN_FOR_SPANS_ENABLED environment variable to true.DeleteItem spans.PutItem spans. Table Primary Keys need to be provided with the ddtrace.config.botocore.dynamodb_primary_key_names_for_tables option or the DD_BOTOCORE_DYNAMODB_TABLE_PRIMARY_KEYS environment variable.UpdateItem spans.CompleteMultipartUpload spans.DD_TRACE_HTTP_CLIENT_ERROR_STATUSES environment variable to configure the list of HTTP status codes that should be considered errors when instrumenting HTTP servers.DD_TRACE_HTTP_CLIENT_TAG_QUERY_STRING instead of ddtrace.config.http_tag_query_stringDD_TRACE_HEADER_TAGS instead of ddtrace.config.trace_http_header_tagsDD_TRACE_REPORT_HOSTNAME instead of ddtrace.config.report_hostnameDD_TRACE_HEALTH_METRICS_ENABLED instead of ddtrace.config.health_metrics_enabledDD_TRACE_ANALYTICS_ENABLED instead of ddtrace.config.analytics_enabledDD_TRACE_CLIENT_IP_HEADER instead of ddtrace.config.client_ip_headerDD_TRACE_CLIENT_IP_ENABLED instead of ddtrace.config.retrieve_client_ipDD_TRACE_PROPAGATION_HTTP_BAGGAGE_ENABLED instead of ddtrace.config.propagation_http_baggage_enabledDD_TRACE_OBFUSCATION_QUERY_STRING_REGEXP to an empty string instead of setting ddtrace.config.global_query_string_obfuscation_disabled to True (default value is False)DD_TRACE_METHODS instead of ddtrace.config.trace_methodsDD_CIVISIBILITY_LOG_LEVEL instead of ddtrace.config.ci_visibility_log_levelDD_TRACE_SAMPLE_RATE environment variable. It will be removed in 3.0.0. Use DD_TRACE_SAMPLING_RULES to configure sampling rates instead.DD_TRACE_API_VERSION=v0.3 is deprecated. Use v0.4 or v0.5 instead.Code security
_DD_IAST_DEBUG environment variable is enabled or not.re.finditer.LLM Observability
LLMObs.enable() did not patch google_generativeai library.modelId when using cross-region inference.Profiling
DD_PROFILING_STACK_V2_ENABLED is set.DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_STACK_V2_ENABLED, or DD_PROFILING_TIMELINE_ENABLED.DD_PROFILING_STACK_V2_ENABLED.Tracing
resource_name.AGENT_CONFIG remote configuration product.celery.apply spans didn't close if the after_task_publish or task_postrun signals didn't get sent when using apply_async, which can happen if there is an internal exception during the handling of the task. This update also marks the span as an error if an exception occurs.celery.apply spans using task_protocol 1 didn't close by improving the check for the task id in the body.http.url span tag contains the full query string when DD_TRACE_OBFUSCATION_QUERY_STRING_REGEXP is set to an empty string.DD_TRACE_RATE_LIMIT environment variable is only applied to spans for which tracer sampling is configured. For spans not matching sampling rules default rate limits should be applied by the Datadog Agent.Other
Nothing published for this version
Nothing published for this version
Nothing published for this version
Add googlecloudsdk and google auth to the Code Security deny list.
Code Security:
Profiling:
Fixes an issue where cpu-time was not profiled for services using gunicorn, when DD_PROFILING_STACK_V2_ENABLED was set.
Fixes an issue where the sample pool could deadlock after fork() by clearing it in the child process.
Fixes an issue where enabling native exporter via DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_TIMELINE_ENABLED or DD_PROFILING_STACK_V2_ENABLED tu
Profiling
DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_TIMELINE_ENABLED or DD_PROFILING_STACK_V2_ENABLED turned off live heap profiling.ensure_binary_or_empty() function, on Python versions before 3.12, with DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.Span. The mapping is populated and used when DD_PROFILING_ENDPOINT_COLLECTION_ENABLED and DD_PROFILING_STACK_V2_ENABLED were set to enable grouping of profiles for endpoints.DD_PROFILING_STACK_V2_ENABLED is set.Tracing
NoneType instead of expected Pin object would throw an error in TracedTopology method.LLM Observability: This fix resolves an issue where LLMObs.enable() did not patch google\_generativeai library.
CODEOWNERS would incorrectly fail to discard line-level trailing comments (eg: @code/owner # my comment would result in codeowners being parsed as @code/owner, #, my, and comment)git binary or .git directory)google.cloud.storage.batch module would fail raising an ImportErrorTypeError exception when it is given a Span with None span_type.Ensures IAST propagation does not raise side effects related to re.finditer.
re.finditer.modelId when using cross-region inference.AGENT_CONFIG remote configuration product.celery: Fixes an issue where celery.apply spans didn't close if the after_task_publish or task_postrun signals didn't get sent when using apply_async,
Tracing
celery.apply spans didn't close if the after_task_publish or task_postrun signals didn't get sent when using apply_async, which can happen if there is an internal exception during the handling of the task. This update also marks the span as an error if an exception occurs.celery.apply spans using task_protocol 1 didn't close by improving the check for the task id in the body.Profiling
DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_STACK_V2_ENABLED, or DD_PROFILING_TIMELINE_ENABLED.DD_PROFILING_STACK_V2_ENABLED.Code Security (IAST): Always report a telemetry log error when an IAST propagation error raises, regardless of whether the _DD_IAST_DEBUG environment
_DD_IAST_DEBUG environment variable is enabled or not.DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.Deprecates the DD_TRACE_SPAN_AGGREGATOR_RLOCK environment variable. It will be removed in v3.0.0.
DD_TRACE_SPAN_AGGREGATOR_RLOCK environment variable. It will be removed in v3.0.0.DD_HTTP_CLIENT_TAG_QUERY_STRING configuration is deprecated and will be removed in v3.0.0. Use DD_TRACE_HTTP_CLIENT_TAG_QUERY_STRING instead.DSM
LLM Observability
model_provider values. Span names are now prefixed with the OpenAI client name (possible values: OpenAI/AzureOpenAI) instead of the default openai prefix to better differentiate whether the request was made to Azure OpenAI or OpenAI. The model_provider field also now corresponds to openai or azure_openai based on the OpenAI client.stream_options={"include_usage": True} option is set on the completion or chat completion call.LLMObs.annotation_context() context manager method, which allows modifying the tags of integration generated LLM Observability spans created while the context manager is active.LLMObs.annotate(prompt={...}) for LLM span kinds. For more information on prompt annotations, see the docs.generate_content calls.openai.request.client tag with the possible values OpenAI/AzureOpenAI to help differentiate whether the request was made to Azure OpenAI or OpenAI.stream_options={"include_usage": True} option is set on the completion or chat completion call.Profiling
asyncio.Lock usages with with context managers.Other
Code Security (ASM)
AttributeError exceptions were being caught, interfering with the proper application cycle.LLM Observability
LLMObs.enable(agentless_enabled=True) or setting DD_LLMOBS_AGENTLESS_ENABLED=1.LLMObs.annotate() method where non-JSON serializable arguments were discarded entirely. Now, the LLMObs.annotate() method safely handles non-JSON-serializable arguments by defaulting to a placeholder text.TypeError in the OpenAI, LangChain, Bedrock, and Anthropic integrations.TypeError. The Anthropic integration now safely tags non-JSON serializable arguments with a default placeholder text.TypeError. The LangChain integration now safely tags non-JSON serializable arguments with a default placeholder text.Other
pymongo.MongoClient can be patch after pymongo is importedNothing published for this version
Fixes a bug where CODEOWNERS would incorrectly fail to discard line-level trailing comments (eg: @code/owner # my comment would result in codeowners b
CI Visibility
CODEOWNERS would incorrectly fail to discard line-level trailing comments (eg: @code/owner # my comment would result in codeowners being parsed as @code/owner, #, my, and comment)git binary or .git directory)Code security
google.cloud.storage.batch module would fail raising an ImportErrorLLM Observability
LLMObs.annotate.Profiling
TypeError exception when it is given a Span with None span_type.Tracing
elasticsearch: Resolves an issue where span tags were not fully populated on "sampled" spans, causing metric dimensions to be incorrect when spans were prematurely marked as sampled, including resource_name.Ensures IAST propagation does not raise side effects related to re.finditer.
re.finditer.modelId when using cross-region inference.AGENT_CONFIG remote configuration product.Always report a telemetry log error when an IAST propagation error raises, regardless of whether the _DD_IAST_DEBUG environment variable is enabled or
Code Security (IAST)
_DD_IAST_DEBUG environment variable is enabled or not.Profiling:
DD_PROFILING_STACK_V2_ENABLED is set.DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_STACK_V2_ENABLED, or DD_PROFILING_TIMELINE_ENABLED.DD_PROFILING_STACK_V2_ENABLED.Tracing
celery.apply spans didn't close if the after_task_publish or task_postrun signals didn't get sent when using apply_async, which can happen if there is an internal exception during the handling of the task. This update also marks the span as an error if an exception occurs.celery.apply spans using task_protocol 1 didn't close by improving the check for the task id in the body.tracing: All public patch modules are deprecated. The non-deprecated methods are included in the __all__ attribute.
set_exc_info.similarity_search spans to LLM Observability as retrieval spans.ml_app and timestamp_ms fields in LLMObs.submit_evaluationmodel tag for openai integration metrics for consistency with the OpenAI SaaS Integration. It has the same value as openai.request.model.__all__ attribute.priority_sampling argument in ddtrace.tracer.Tracer.configure(...).attrs installed by default on some Ubuntu installations was treated as incompatible with library injectionTypeError. The Anthropic integration now safely tags non-JSON serializable arguments with a default placeholder text..git was incorrectly being stripped from repository URLs when extracting service names, resulting in g, i, or t being removed (eg: test-environment.git incorrectly becoming test-environmen)start_execution commands. This re-enables distributed tracing when a Python service invokes a properly instrumented Step Function.celery or gunicorn workers. The LLM Observability writer thread now automatically restarts when a forked process is detected.TypeError exception at runtime.X-Datadog-Trace-Count payload header.ddtrace package files were published with incorrect file attributes.session_id was being defaulted to trace_id, which was causing unexpected UI behavior.asyncio.TimeoutErrors were not being propagated correctly from canceled OpenAI API requests.DD_PROFILING_TAGS and DD_TAGS to the libdatadog exporter, a new exporter codepath which is enabled when either one of the following is set, DD_PROFILING_STACK_V2_ENABLED, DD_PROFILING_EXPORT_LIBDD_ENABLED, or DD_PROFILING_TIMELINE_ENABLED or dd-trace-py is running in an injected environment.DD_PRIORITY_SAMPLING configuration option. This option is not used in any ddtrace>=2.0 releases.Nothing published for this version
Fixes an issue where stack v2 couldn't be enabled as pthread was not properly linked on some debian based images for aarch64 architecture.
AGENT_CONFIG remote configuration product.This fix resolves an issue where exploit prevention was not properly blocking requests with custom redirection actions.
Code Security
Initializer object is always reset and freed before the Python runtime.LLM Observability
LLMObs.enable(agentless_enabled=True) or setting DD_LLMOBS_AGENTLESS_ENABLED=1.LLMObs.annotate() method where non-JSON serializable arguments were discarded entirely. Now, the LLMObs.annotate() method safely handles non-JSON-serializable arguments by defaulting to a placeholder text.TypeError in the OpenAI, LangChain, Bedrock, and Anthropic integrations.TypeError. The Anthropic integration now safely tags non-JSON serializable arguments with a default placeholder text.TypeError. The LangChain integration now safely tags non-JSON serializable arguments with a default placeholder text.Profiling
DD_PROFILING_STACK_V2_ENABLED is set.DD_PROFILING_EXPORT_LIBDD_ENABLED or DD_PROFILING_TIMELINE_ENABLED.DD_PROFILING_EXPORT_LIBDD_ENABLED, DD_PROFILING_STACK_V2_ENABLED, or DD_PROFILING_TIMELINE_ENABLED.DD_PROFILING_STACK_V2_ENABLED.Tracing
celery.apply spans didn't close if the after_task_publish or task_postrun signals didn't get sent when using apply_async, which can happen if there is an internal exception during the handling of the task. This update also marks the span as an error if an exception occurs.celery.apply spans using task_protocol 1 didn't close by improving the check for the task id in the body.TypeError exception would be raised if the first message's topic() returned None during consumption.library injection: Resolves an issue where the version of attrs installed by default on some Ubuntu installations was treated as incompatible with lib
attrs installed by default on some Ubuntu installations was treated as incompatible with library injectionAttributeError exceptions were being caught, interfering with the proper application cycle.SSI: This fix ensures injection denylist is included in published OCI package.
config: DD_TRACE_128_BIT_TRACEID_LOGGING_ENABLED is deprecated. Trace id logging format is now configured automatically.
model tag for openai integration metrics for consistency with the OpenAI SaaS Integration. It has the same value as openai.request.model.server.address tag to all <database>.query spans (ex: postgres.query). This tag stores the name of the database host.ddtrace.trace.Context to the public api. This class can now be used to propagate context across execution boundaries (ex: threads).DD_TRACE_128_BIT_TRACEID_LOGGING_ENABLED is deprecated. Trace id logging format is now configured automatically.ddtrace.contrib.[integration_name] package. Use attributes exposed in ddtrace.contrib.[integration_name].__all__ instead. The following are impacted:
aioredis, algoliasearch. anthropic, aredis, asgi, asyncpg, aws_lambda, boto, botocore, bottle, cassandra, celery, cherrypy, consul, coverage, django, dogpile_cache, dramatiq, elasticsearch, falcon, fastapi, flask, flask_cache, futures, gevent, graphql, grpc, httplib, httpx, jinja2, kafka, kombu, langchain, logbook, logging, loguru, mako, mariadb, molten, mongoengine, mysql, mysqldb, openai, psycopg, pylibmc, pymemcache, pymongo, pymysql, pynamodb, pyodbc, pyramid, redis, rediscluster, requests, sanic, snowflake, sqlalchemy, sqlite3, starlette, structlog, subprocess, tornado, urllib, urllib3, vertica, webbrowser, wsgi, yaaredisCI Visibility: Resolves an issue where exceptions other than timeouts and connection errors raised while fetching the list of skippable tests for ITR were not being handled correctly and caused the tracer to crash.
CI Visibility: Fixes a bug where .git was incorrectly being stripped from repository URLs when extracting service names, resulting in g, i, or t being removed (eg: test-environment.git incorrectly becoming test-environmen)
LLM Observability: Resolves an issue where custom trace filters were being overwritten in forked processes.
tracing: Fixes a side-effect issue with module import callbacks that could cause a runtime exception.
LLM Observability: Resolves an issue where session_id was being defaulted to trace_id, which was causing unexpected UI behavior.
LLM Observability: Resolves an issue where LLM Observability spans were not being submitted in forked processes, such as when using celery or gunicorn workers. The LLM Observability writer thread now automatically restarts when a forked process is detected.
tracing: Fixes an issue with some module imports with native specs that don't support attribute assignments, resulting in a TypeError exception at runtime.
tracing: Resolves an issue where ddtrace package files were published with incorrect file attributes.
tracing: Resolves an issue where django db instrumentation could fail.
openai: Fixes a bug where asyncio.TimeoutErrors were not being propagated correctly from canceled OpenAI API requests.
aiobotocore: Fixes an issue where the _make_api_call arguments were not captured correctly when using keyword arguments.
tracing(django): Resolves a bug where ddtrace was exhausting a Django stream response before returning it to user.
LLM Observability: Fixes an issue in the OpenAI integration where integration metrics would still be submitted even if LLMObs.enable(agentless_enabled=True) was set.
internal: Fixes the Already mutably borrowed error when rate limiter is accessed across threads.
internal: Fixes the Already mutably borrowed error by reverting back to pure-python rate limiter.
Code Security: Adds null pointer checks when creating new objects ids.
profiling: Fixes an issue where the profiler could erroneously try to load protobuf in autoinjected environments, where it is not available.
crashtracking: Fixes an issue where crashtracking environment variables for Python were inconsistent with those used by other runtimes.
profiling: Fixes endpoint profiling for stack v2 when DD_PROFILING_STACK_V2_ENABLED is set.
profiling: Turns on the new native exporter when DD_PROFILING_TIMELINE_ENABLED=True is set.
Nothing published for this version
One column per quarter.