Revert changes as they were centred around adding the request body, which we are no longer doing. Add test coverage to ensure span streaming is working as expected

Author: ericapisani

sentry_sdk/integrations/_wsgi_common.py

@@ -209,15 +209,6 @@ def _is_json_content_type(ct: "Optional[str]") -> bool:
     )
 
 
-def _serialize_request_body_data(data: "Any") -> str:
-    def _default(value: "Any") -> "Any":
-        if isinstance(value, AnnotatedValue):
-            return value.value
-        return str(value)
-
-    return json.dumps(data, default=_default)
-
-
 def _filter_headers(
     headers: "Mapping[str, str]",
     use_annotated_value: bool = True,

sentry_sdk/integrations/flask.py

@@ -5,16 +5,11 @@
 from sentry_sdk.integrations._wsgi_common import (
     DEFAULT_HTTP_METHODS_TO_CAPTURE,
     RequestExtractor,
-    _serialize_request_body_data,
-    request_body_within_bounds,
 )
 from sentry_sdk.integrations.wsgi import SentryWsgiMiddleware
 from sentry_sdk.scope import should_send_default_pii
-from sentry_sdk.traces import StreamedSpan, _get_current_streamed_span
 from sentry_sdk.tracing import SOURCE_FOR_STYLE
-from sentry_sdk.tracing_utils import has_span_streaming_enabled
 from sentry_sdk.utils import (
-    AnnotatedValue,
     capture_internal_exceptions,
     ensure_integration_enabled,
     event_from_exception,
@@ -41,11 +36,9 @@
     from flask.signals import (
         before_render_template,
         got_request_exception,
-        request_finished,
         request_started,
     )
     from markupsafe import Markup
-    from werkzeug.exceptions import ClientDisconnected
 except ImportError:
     raise DidNotEnable("Flask is not installed")
 
@@ -95,7 +88,6 @@ def setup_once() -> None:
 
         before_render_template.connect(_add_sentry_trace)
         request_started.connect(_request_started)
-        request_finished.connect(_request_finished)
         got_request_exception.connect(_capture_exception)
 
         old_app = Flask.__call__
@@ -167,72 +159,6 @@ def _request_started(app: "Flask", **kwargs: "Any") -> None:
     scope.add_event_processor(evt_processor)
 
 
-def _request_finished(sender: "Flask", response: "Any", **kwargs: "Any") -> None:
-    integration = sentry_sdk.get_client().get_integration(FlaskIntegration)
-    if integration is None:
-        return
-
-    client = sentry_sdk.get_client()
-    if has_span_streaming_enabled(client.options):
-        request = flask_request._get_current_object()
-        _set_request_body_data_on_streaming_segment(request, client)
-
-
-def _set_request_body_data_on_streaming_segment(
-    request: "Request", client: "sentry_sdk.client.BaseClient"
-) -> None:
-    current_span = _get_current_streamed_span()
-    if type(current_span) is not StreamedSpan:
-        return
-
-    with capture_internal_exceptions():
-        content_length = int(request.content_length or 0)
-
-        # Proceeding without a content length means that we may be consuming the request
-        # without respecting the bounds specified by the user via `max_request_body_size`
-        # option in the SDK.
-        if not content_length:
-            return
-
-        if not request_body_within_bounds(client, content_length):
-            data = AnnotatedValue.substituted_because_over_size_limit()
-        else:
-            raw_data = getattr(request, "_cached_data", None)
-            parsed_body = None
-            if "form" in request.__dict__:
-                extractor = FlaskRequestExtractor(request)
-                parsed_body = extractor.parsed_body()
-            elif raw_data is not None:
-                extractor = FlaskRequestExtractor(request)
-                if extractor.is_json():
-                    parsed_body = extractor.json()
-            else:
-                # The route never read the body via Werkzeug, but it
-                # may have consumed wsgi.input directly. get_data()
-                # raises ClientDisconnected if the stream is exhausted.
-                try:
-                    raw_data = request.get_data()
-                except ClientDisconnected:
-                    raw_data = None
-
-                if raw_data:
-                    extractor = FlaskRequestExtractor(request)
-                    if extractor.is_json():
-                        parsed_body = extractor.json()
-
-            if parsed_body is not None:
-                data = parsed_body
-            elif raw_data:
-                data = AnnotatedValue.substituted_because_raw_data()
-            else:
-                return
-
-        current_span._segment.set_attribute(
-            "http.request.body.data",
-            _serialize_request_body_data(data),
-        )
-
-
 class FlaskRequestExtractor(RequestExtractor):
     def env(self) -> "Dict[str, str]":
         return self.request.environ

sentry_sdk/integrations/starlette.py

@@ -1,4 +1,5 @@
 import functools
+import json
 import sys
 import warnings
 from collections.abc import Set
@@ -17,7 +18,6 @@
     DEFAULT_HTTP_METHODS_TO_CAPTURE,
     HttpCodeRangeContainer,
     _is_json_content_type,
-    _serialize_request_body_data,
     request_body_within_bounds,
 )
 from sentry_sdk.integrations.asgi import SentryAsgiMiddleware
@@ -241,6 +241,16 @@ async def _sentry_send(*args: "Any", **kwargs: "Any") -> "Any":
     return middleware_class
 
 
+def _serialize_request_body_data(data: "Any") -> str:
+    # data may be a JSON-serializable value, an AnnotatedValue, or a dict with AnnotatedValue values
+    def _default(value: "Any") -> "Any":
+        if isinstance(value, AnnotatedValue):
+            return value.value
+        return str(value)
+
+    return json.dumps(data, default=_default)
+
+
 def _set_request_body_data_on_streaming_segment(
     info: "Optional[Dict[str, Any]]",
 ) -> None: