Merge branch 'main' into evan.li/release-ragas

.github/CODEOWNERS

@@ -166,12 +166,14 @@ ddtrace/internal/remoteconfig       @DataDog/remote-config @DataDog/apm-core-pyt
 tests/internal/remoteconfig         @DataDog/remote-config @DataDog/apm-core-python
 
 # API SDK
+ddtrace/trace/                                    @DataDog/apm-sdk-api-python
 ddtrace/_trace/                                    @DataDog/apm-sdk-api-python
 ddtrace/opentelemetry/                             @DataDog/apm-sdk-api-python
 ddtrace/internal/opentelemetry                     @DataDog/apm-sdk-api-python
 ddtrace/opentracer/                                @DataDog/apm-sdk-api-python
 ddtrace/propagation/                               @DataDog/apm-sdk-api-python
 ddtrace/filters.py                                 @DataDog/apm-sdk-api-python
+ddtrace/provider.py                                @DataDog/apm-sdk-api-python
 ddtrace/pin.py                                     @DataDog/apm-sdk-api-python
 ddtrace/sampler.py                                 @DataDog/apm-sdk-api-python
 ddtrace/sampling_rule.py                           @DataDog/apm-sdk-api-python

.github/workflows/build_python_3.yml

@@ -35,7 +35,8 @@ jobs:
               cibuildwheel --print-build-identifiers --platform linux --arch x86_64,i686 | jq -cR '{only: ., os: "ubuntu-latest"}' \
               && cibuildwheel --print-build-identifiers --platform linux --arch aarch64  | jq -cR '{only: ., os: "arm-4core-linux"}' \
               && cibuildwheel --print-build-identifiers --platform windows --arch AMD64,x86 | grep -v 313 | jq -cR '{only: ., os: "windows-latest"}' \
-              && cibuildwheel --print-build-identifiers --platform macos --arch x86_64,universal2 | jq -cR '{only: ., os: "macos-13"}'
+              && cibuildwheel --print-build-identifiers --platform macos --arch x86_64 | jq -cR '{only: ., os: "macos-13"}' \
+              && cibuildwheel --print-build-identifiers --platform macos --arch arm64 | jq -cR '{only: ., os: "macos-latest"}'
             } | jq -sc
           )
           echo $MATRIX_INCLUDE
@@ -112,6 +113,7 @@ jobs:
             choco install -y 7zip &&
             7z d -r "{wheel}" *.c *.cpp *.cc *.h *.hpp *.pyx &&
             move "{wheel}" "{dest_dir}"
+          CIBW_TEST_COMMAND: "python {project}/tests/smoke_test.py"
           # DEV: Uncomment to debug MacOS
           # CIBW_BUILD_VERBOSITY_MACOS: 3
 
@@ -152,6 +154,7 @@ jobs:
             choco install -y 7zip &&
             7z d -r "{wheel}" *.c *.cpp *.cc *.h *.hpp *.pyx &&
             move "{wheel}" "{dest_dir}"
+          CIBW_TEST_COMMAND: "python {project}/tests/smoke_test.py"
           # DEV: Uncomment to debug MacOS
           # CIBW_BUILD_VERBOSITY_MACOS: 3
 

.github/workflows/check_safe_main_merge.yml

@@ -0,0 +1,27 @@
+name: Check for Safe main Merge
+
+on:
+  pull_request:
+    branches:
+      - '3.x-staging'
+
+jobs:
+  check-merge:
+    runs-on: ubuntu-latest
+    steps:
+      # Step 1: Checkout the repository
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      # Step 2: Fetch the main branch
+      - name: Fetch main branch
+        run: git fetch origin main
+
+      # Step 3: Attempt to merge
+      - name: Check merge conflicts
+        run: |
+          git merge --no-commit --no-ff origin/main || exit 1
+      # Step 4: Clean up the merge (optional)
+      - name: Abort merge
+        if: failure()
+        run: git merge --abort

.gitlab/benchmarks.yml

@@ -96,6 +96,7 @@ benchmark-serverless:
   tags: ["arch:amd64"]
   when: on_success
   needs: [ "benchmark-serverless-trigger" ]
+  allow_failure: true
   script:
     - git clone https://gitlab-ci-token:${CI_JOB_TOKEN}@gitlab.ddbuild.io/DataDog/serverless-tools.git ./serverless-tools && cd ./serverless-tools
     - ./ci/check_trigger_status.sh

CHANGELOG.md

@@ -4,6 +4,21 @@ Changelogs for versions not listed here can be found at https://github.com/DataD
 
 ---
 
+## 2.19.2
+### Bug Fixes
+
+- Tracing
+  - celery: Fixes an issue where `celery.apply` spans from Celery prerun got closed too soon leading to span tags being missing.
+  - openai: Fixes a patching issue where asynchronous moderation endpoint calls resulted in coroutine scheduling errors.
+  - openai: Ensures the OpenAI integration is compatible with Python versions 3.12 and 3.13.
+  - vertexai: Resolves an issue with `chat.send_message()` where the content keyword argument was not parsed correctly.
+- LLM Observability
+  - This fix resolves an issue where annotating a span with non latin-1 (but valid utf-8) input/output values resulted in encoding errors.
+- Lib-Injection
+  - Fixes incorrect telemetry data payload format.
+
+---
+
 ## 2.19.1
 ### Bug Fixes
 

ddtrace/_trace/trace_handlers.py

@@ -109,11 +109,14 @@ def _get_parameters_for_new_span_directly_from_context(ctx: core.ExecutionContex
 def _start_span(ctx: core.ExecutionContext, call_trace: bool = True, **kwargs) -> "Span":
     span_kwargs = _get_parameters_for_new_span_directly_from_context(ctx)
     call_trace = ctx.get_item("call_trace", call_trace)
-    tracer = (ctx.get_item("middleware") or ctx["pin"]).tracer
+    tracer = ctx.get_item("tracer") or (ctx.get_item("middleware") or ctx["pin"]).tracer
     distributed_headers_config = ctx.get_item("distributed_headers_config")
     if distributed_headers_config:
         trace_utils.activate_distributed_headers(
-            tracer, int_config=distributed_headers_config, request_headers=ctx["distributed_headers"]
+            tracer,
+            int_config=distributed_headers_config,
+            request_headers=ctx["distributed_headers"],
+            override=ctx.get_item("distributed_headers_config_override"),
         )
     distributed_context = ctx.get_item("distributed_context")
     if distributed_context and not call_trace:
@@ -126,6 +129,42 @@ def _start_span(ctx: core.ExecutionContext, call_trace: bool = True, **kwargs) -
     return span
 
 
+def _set_web_frameworks_tags(ctx, span, int_config):
+    span.set_tag_str(COMPONENT, int_config.integration_name)
+    span.set_tag_str(SPAN_KIND, SpanKind.SERVER)
+    span.set_tag(_SPAN_MEASURED_KEY)
+
+    analytics_enabled = ctx.get_item("analytics_enabled")
+    analytics_sample_rate = ctx.get_item("analytics_sample_rate", True)
+
+    # Configure trace search sample rate
+    if (config._analytics_enabled and analytics_enabled is not False) or analytics_enabled is True:
+        span.set_tag(_ANALYTICS_SAMPLE_RATE_KEY, analytics_sample_rate)
+
+
+def _on_web_framework_start_request(ctx, int_config):
+    request_span = ctx.get_item("req_span")
+    _set_web_frameworks_tags(ctx, request_span, int_config)
+
+
+def _on_web_framework_finish_request(
+    span, int_config, method, url, status_code, query, req_headers, res_headers, route, finish
+):
+    trace_utils.set_http_meta(
+        span=span,
+        integration_config=int_config,
+        method=method,
+        url=url,
+        status_code=status_code,
+        query=query,
+        request_headers=req_headers,
+        response_headers=res_headers,
+        route=route,
+    )
+    if finish:
+        span.finish()
+
+
 def _on_traced_request_context_started_flask(ctx):
     current_span = ctx["pin"].tracer.current_span()
     if not ctx["pin"].enabled or not current_span:
@@ -761,6 +800,10 @@ def listen():
     core.on("azure.functions.request_call_modifier", _on_azure_functions_request_span_modifier)
     core.on("azure.functions.start_response", _on_azure_functions_start_response)
 
+    # web frameworks general handlers
+    core.on("web.request.start", _on_web_framework_start_request)
+    core.on("web.request.finish", _on_web_framework_finish_request)
+
     core.on("test_visibility.enable", _on_test_visibility_enable)
     core.on("test_visibility.disable", _on_test_visibility_disable)
     core.on("test_visibility.is_enabled", _on_test_visibility_is_enabled, "is_enabled")
@@ -769,6 +812,14 @@ def listen():
     core.on("rq.queue.enqueue_job", _propagate_context)
 
     for context_name in (
+        # web frameworks
+        "aiohttp.request",
+        "bottle.request",
+        "cherrypy.request",
+        "falcon.request",
+        "molten.request",
+        "pyramid.request",
+        "sanic.request",
         "flask.call",
         "flask.jsonify",
         "flask.render_template",
@@ -779,6 +830,7 @@ def listen():
         "django.template.render",
         "django.process_exception",
         "django.func.wrapped",
+        # non web frameworks
         "botocore.instrumented_api_call",
         "botocore.instrumented_lib_function",
         "botocore.patched_kinesis_api_call",

ddtrace/appsec/_ddwaf/ddwaf_types.py

@@ -22,9 +22,12 @@
 
 if system() == "Linux":
     try:
+        asm_config._bypass_instrumentation_for_waf = True
         ctypes.CDLL(ctypes.util.find_library("rt"), mode=ctypes.RTLD_GLOBAL)
     except Exception:  # nosec
         pass
+    finally:
+        asm_config._bypass_instrumentation_for_waf = False
 
 ARCHI = machine().lower()
 

ddtrace/appsec/_processor.py

@@ -4,13 +4,19 @@
 from json.decoder import JSONDecodeError
 import os
 import os.path
+from typing import TYPE_CHECKING
 from typing import Any
 from typing import Dict
 from typing import List
 from typing import Optional
 from typing import Set
 from typing import Tuple
 from typing import Union
+
+
+if TYPE_CHECKING:
+    import ddtrace.appsec._ddwaf as ddwaf
+
 import weakref
 
 from ddtrace._trace.processor import SpanProcessor
@@ -167,14 +173,17 @@ def __post_init__(self) -> None:
     def delayed_init(self) -> None:
         try:
             if self._rules is not None and not hasattr(self, "_ddwaf"):
-                self._ddwaf = ddwaf.DDWaf(
+                from ddtrace.appsec._ddwaf import DDWaf  # noqa: E402
+                import ddtrace.appsec._metrics as metrics  # noqa: E402
+
+                self.metrics = metrics
+                self._ddwaf = DDWaf(
                     self._rules, self.obfuscation_parameter_key_regexp, self.obfuscation_parameter_value_regexp
                 )
-            _set_waf_init_metric(self._ddwaf.info)
+                self.metrics._set_waf_init_metric(self._ddwaf.info)
         except Exception:
             # Partial of DDAS-0005-00
             log.warning("[DDAS-0005-00] WAF initialization failed")
-            raise
         self._update_required()
 
     def _update_required(self):
@@ -193,7 +202,7 @@ def _update_rules(self, new_rules: Dict[str, Any]) -> bool:
         if asm_config._asm_static_rule_file is not None:
             return result
         result = self._ddwaf.update_rules(new_rules)
-        _set_waf_updates_metric(self._ddwaf.info)
+        self.metrics._set_waf_updates_metric(self._ddwaf.info)
         self._update_required()
         return result
 
@@ -241,7 +250,7 @@ def waf_callable(custom_data=None, **kwargs):
             return self._waf_action(span._local_root or span, ctx, custom_data, **kwargs)
 
         _asm_request_context.set_waf_callback(waf_callable)
-        _asm_request_context.add_context_callback(_set_waf_request_metrics)
+        _asm_request_context.add_context_callback(self.metrics._set_waf_request_metrics)
         if headers is not None:
             _asm_request_context.set_waf_address(SPAN_DATA_NAMES.REQUEST_HEADERS_NO_COOKIES, headers)
             _asm_request_context.set_waf_address(
@@ -436,10 +445,3 @@ def on_span_finish(self, span: Span) -> None:
                     del self._span_to_waf_ctx[s]
                 except Exception:  # nosec B110
                     pass
-
-
-# load waf at the end only to avoid possible circular imports with gevent
-import ddtrace.appsec._ddwaf as ddwaf  # noqa: E402
-from ddtrace.appsec._metrics import _set_waf_init_metric  # noqa: E402
-from ddtrace.appsec._metrics import _set_waf_request_metrics  # noqa: E402
-from ddtrace.appsec._metrics import _set_waf_updates_metric  # noqa: E402