3.14.2

Estimated end-of-life date, accurate to within three months: 09-2026
See the support level definitions for more information.

Bug Fixes

• sampling: This change prevents the DatadogSampler from getting recreated whenever the SpanAggregator is reset, and
  instead updates the rate limiter that the sampler uses.

• dynamic instrumentation: fix an issue that prevented multiple probes on the same location from being instrumented.

• source code integration: check that DD_GIT_COMMIT_SHA and DD_GIT_REPOSITORY_URL are defined before using the git command.

3.13.3

Estimated end-of-life date, accurate to within three months: 09-2026
See the support level definitions for more information.

Bug Fixes

• sampling: This change prevents the DatadogSampler from getting recreated whenever the SpanAggregator is reset, and instead updates the rate limiter that the sampler uses.

• source code integration: check that DD_GIT_COMMIT_SHA and DD_GIT_REPOSITORY_URL are defined before using the git command.

• CI Visibility: This fix resolves an issue where coverage from sessions with pytest-xdist were not submitted with the proper session id, preventing Test Impact Analysis feature from working properly.

3.12.8

Estimated end-of-life date, accurate to within three months: 09-2026
See the support level definitions for more information.

Bug Fixes

• exception replay: fixed an issue that prevented snapshots from retrieving local variables from traceback frames from exception thrown by Celery tasks.
• source code integration: check that DD_GIT_COMMIT_SHA and DD_GIT_REPOSITORY_URL are defined before using the git command.

3.14.1

Estimated end-of-life date, accurate to within three months: 08-2026
See the support level definitions for more information.

Bug Fixes

• AAP:
  • This fix disables grpc threat monitoring, as it could generate false positives.

3.14.0

Estimated end-of-life date, accurate to within three months: 08-2026
See the support level definitions for more information.

New Features

• ai_guard
  • add AI Guard evaluations support to LangChain instrumentation
  • add evaluation support to streaming LangChain APIs
• LLM Observability
  • ml_app is now optional, defaulting to service. while it is still recommended to set ml_app, enabling LLM Observability will no longer throw if one is not provided or propagated from an upstream service.
  • This introduces a tool_definitions parameter to the LLMObs.annotate() method for tool calling scenarios. Users can now pass a list of tool definition dictionaries directly to annotate LLM spans with available tools. Each tool definition must include a name (string) field, with optional description (string) and schema (JSON-serializable dictionary) fields.
  • Experiments (beta) dataset methods (create_dataset, create_dataset_from_csv and pull_dataset) argument name is changed to dataset_name.

Bug Fixes

• AAP
  • make sure the status code for downstream requests is properly sent to libddwaf.
  • Resolves an incompatibility with gevent>=25.8.1 that would cause a deadlock when starting the waf via remote config.
• django: Fixes issue causing ValueError: coroutine already executing on Python 3.13+ with django.utils.decorators.async_only_middleware.
• asyncpg: Fix the error "Error: expected pool connect callback to return an instance of 'asyncpg.connection.Connection', got 'ddtrace.contrib.internal.asyncpg.patch._TracedConnection'" due to using the custom connect option. With this fix, postgres.connect spans will be created when this option is used.
• exception replay: fixed an issue that prevented snapshots from retrieving local variables from traceback frames from exception thrown by Celery tasks.
• LLM Observability
  • This fix resolves an issue where decorated functions returning responses with an ambiguous truth value (e.g. pandas dataframes) would raise an error due to the inability to coerce the value into a boolean.
  • Fixes an issue where certain Google GenAI LLM requests were not being traced due to importing from google.genai.types on startup.
• Fixed an issue that could have caused some products to fail to start properly in applications that use pkg_resources, either directly or indirectly.
• profiling: Upgrades echion to pick up critical bug fixes and performance improvements.
• CI Visibility: This fix resolves an issue where coverage from sessions with pytest-xdist were not submitted with the proper session id, preventing Test Impact Analysis feature from working properly.

3.14.0rc2

Estimated end-of-life date, accurate to within three months: 08-2026
See the support level definitions for more information.

New Features

• ai_guard
  • add AI Guard evaluations support to LangChain instrumentation
  • add evaluation support to streaming LangChain APIs
• LLM Observability
  • ml_app is now optional, defaulting to service. while it is still recommended to set ml_app, enabling LLM Observability will no longer throw if one is not provided or propagated from an upstream service.
  • This introduces a tool_definitions parameter to the LLMObs.annotate() method for tool calling scenarios. Users can now pass a list of tool definition dictionaries directly to annotate LLM spans with available tools. Each tool definition must include a name (string) field, with optional description (string) and schema (JSON-serializable dictionary) fields.
  • Experiments (beta) dataset methods (create_dataset, create_dataset_from_csv and pull_dataset) argument name is changed to dataset_name.

Bug Fixes

• AAP
  • make sure the status code for downstream requests is properly sent to libddwaf.
  • Resolves an incompatibility with gevent>=25.8.1 that would cause a deadlock when starting the waf via remote config.
• django: Fixes issue causing ValueError: coroutine already executing on Python 3.13+ with django.utils.decorators.async_only_middleware.
• asyncpg: Fix the error "Error: expected pool connect callback to return an instance of 'asyncpg.connection.Connection', got 'ddtrace.contrib.internal.asyncpg.patch._TracedConnection'" due to using the custom connect option. With this fix, postgres.connect spans will be created when this option is used.
• exception replay: fixed an issue that prevented snapshots from retrieving local variables from traceback frames from exception thrown by Celery tasks.
• LLM Observability
  • This fix resolves an issue where decorated functions returning responses with an ambiguous truth value (e.g. pandas dataframes) would raise an error due to the inability to coerce the value into a boolean.
  • Fixes an issue where certain Google GenAI LLM requests were not being traced due to importing from google.genai.types on startup.
• Fixed an issue that could have caused some products to fail to start properly in applications that use pkg_resources, either directly or indirectly.
• profiling: Upgrades echion to pick up critical bug fixes and performance improvements.
• CI Visibility: This fix resolves an issue where coverage from sessions with pytest-xdist were not submitted with the proper session id, preventing Test Impact Analysis feature from working properly.

3.12.7

Estimated end-of-life date, accurate to within three months: 08-2026
See the support level definitions for more information.

Bug Fixes

• tracing:
  • Fixes a bug where calling __repr__ on SpanPointer objected raised an AttributeError. This caused aws_lambdas to crash when debug logging was enabled.

3.13.1

Estimated end-of-life date, accurate to within three months: 08-2026
See the support level definitions for more information.

Bug Fixes

• django:
  • Fixes issue causing ValueError: coroutine already executing on Python 3.13+ with django.utils.decorators.async_only_middleware.

3.14.0rc1

Estimated end-of-life date, accurate to within three months: 08-2026
See the support level definitions for more information.

New Features

• ai_guard: add AI Guard evaluations support to LangChain instrumentation
• LLM Observability
  • ml_app is now optional, defaulting to service. while it is still recommended to set ml_app, enabling LLM Observability will no longer throw if one is not provided or propagated from an upstream service.
  • This introduces a tool_definitions parameter to the LLMObs.annotate() method for tool calling scenarios. Users can now pass a list of tool definition dictionaries directly to annotate LLM spans with available tools. Each tool definition must include a name (string) field, with optional description (string) and schema (JSON-serializable dictionary) fields.

Bug Fixes

• asyncpg: Fix the error "Error: expected pool connect callback to return an instance of 'asyncpg.connection.Connection', got 'ddtrace.contrib.internal.asyncpg.patch._TracedConnection'" due to using the custom connect option. With this fix, postgres.connect spans will be created when this option is used.
• exception replay: fixed an issue that prevented snapshots from retrieving local variables from traceback frames from exception thrown by Celery tasks.
• LLM Observability: This fix resolves an issue where decorated functions returning responses with an ambiguous truth value (e.g. pandas dataframes) would raise an error due to the inability to coerce the value into a boolean.
• Fixed an issue that could have caused some products to fail to start properly in applications that use pkg_resources, either directly or indirectly.
• profiling: Upgrades echion to pick up critical bug fixes and performance improvements.
• CI Visibility: This fix resolves an issue where coverage from sessions with pytest-xdist were not submitted with the proper session id, preventing Test Impact Analysis feature from working properly.

3.13.0

Estimated end-of-life date, accurate to within three months: 08-2026
See the support level definitions for more information.

Deprecation Notes

• tracing: ddtrace.tracer.Pin is deprecated and will be removed in version 4.0.0. To manage configuration of the tracer or integrations please use environment variables.

New Features

• opentelemetry

  • Adds experimental support for exporting logs using the OTLP protocol. To enable this feature install the opentelemetry-otlp-exporter and set DD_LOGS_OTEL_ENABLED=true.
  • Ensures that ddtrace log injection is disabled when OpenTelemetry logs are enabled (DD_LOGS_OTEL_ENABLED=true).

• LLM Observability

  • Adds support for collecting tool definitions, tool calls and tool results in the Amazon Bedrock integration.
  • Adds support for collecting tool definitions, tool calls and tool results in the Google GenAI integration.
  • This introduces support for tool calls and tool results in LLMObs message annotations when using custom instrumentation. The LLMObs.annotate() method now accept input and output data with optional to ol_calls and tool_results fields for function calling scenarios.
  • Adds support for collecting tool definitions, tool calls and tool results in the OpenAI integration.
  • Adds support for collecting tool definitions, tool calls and tool results in the VertexAI integration.

• Django: Added the DD_DJANGO_ALWAYS_CREATE_DATABASE_SPANS config option (default: true).

  When enabled, the Django integration always generates a database span for every operation, even if the underlying database engine is already instrumented. This ensures complete coverage but may produce duplicat
  e spans and extra overhead.

  When disabled, spans are only created if the database engine is not instrumented.

  To avoid multiple spans per database call, we recommend disabling this option. DD_DJANGO_ALWAYS_CREATE_DATABASE_SPANS=false

• AAP

  • This introduces endpoint discovery for Flask applications. It allows the collection of API endpoints of a Flask application at startup.
  • This introduces endpoint discovery for FastAPI applications. It allows the collection of API endpoints of a FastAPI application at startup.

• Bug Fixes

• AAP: Fixes an issue where security signals would be incorrectly reported on an inferred proxy service instead of the current service.

• CI Visibility: This fix resolves an issue where the pytest plugin would hold a reference to test exceptions beyond the end of the test, preventing them from being garbage-collected and increasing memory usage.

• psycopg: This fix resolves a potential circular import with the psycopg3 contrib.

• internal: This fix resolves an issue where the tracer flare was not sent when DD_TRACE_AGENT_URL was not set, as the default URL was not used.

• tracing

  • Fixes issue with websocket.receive span not closing exactly when another websocket.receive span was opened.
  • Fixes duration of websocket handshake span such that the handshake span closes when the connection is upgraded.
  • Fix where websocket.close parent should be the handshake span when configuration is disabled.
  • Resolves an issue where calling ddtrace.trace.tracer.configure(...) resets the trace writer buffer, causing spans to be dropped.

• Code Security: Fixed a crash in the taint-aware modulo aspect when formatting SQLAlchemy objects whose __repr__ can raise (e.g., inside complex CASE expressions).

• LLM Observability: Properly parse DD_TAGS onto LLM Observability span events' tags.

• sampling

  • Fixed a bug where single span sampling rules were completely ignored when stats computation was enabled, causing all single-sampled spans to be dropped. Note that single span sampling is only used to
    override trace sampling decisions to keep a span where the trace will be dropped.
  • Removed automatic setting of local trace sampling rule tag (_dd.p.dm=-3) during distributed tracing header extraction.

• ssi, crashtracker: This fix resolves an issue where crashtracker receiver binary was not available in an injected environment.

Other Changes

• LLM Observability: add ability to override the URL origin.
• openai: Removes I/O tags, including llm prompts, parameters, and token usage, from APM spans for OpenAI audio, image, and moderation requests.