.

Author: sentrivana

sentry_sdk/_span_batcher.py

@@ -13,7 +13,6 @@
 
 
 class SpanBatcher(Batcher["StreamedSpan"]):
-    # TODO[span-first]: adjust flush/drop defaults
     # MAX_BEFORE_FLUSH should be lower than MAX_BEFORE_DROP, so that there is
     # a bit of a buffer for spans that appear between setting the flush event
     # and actually flushing the buffer.

sentry_sdk/scope.py

@@ -1246,9 +1246,8 @@ def start_streamed_span(
 
             if is_ignored_span(name, attributes):
                 return NoOpStreamedSpan(
-                    trace_id=propagation_context.trace_id,
-                    unsampled_reason="ignored",
                     scope=self,
+                    unsampled_reason="ignored",
                 )
 
             sampled, sample_rate, sample_rand, outcome = make_sampling_decision(
@@ -1258,9 +1257,8 @@ def start_streamed_span(
             )
             if sampled is False:
                 return NoOpStreamedSpan(
-                    trace_id=propagation_context.trace_id,
-                    unsampled_reason=outcome,
                     scope=self,
+                    unsampled_reason=outcome,
                 )
 
             if sample_rate is not None:
@@ -1282,11 +1280,13 @@ def start_streamed_span(
 
         # This is a child span; take propagation context from the parent span
         with new_scope():
-            if is_ignored_span(name, attributes) or isinstance(
-                parent_span, NoOpStreamedSpan
-            ):
+            if is_ignored_span(name, attributes):
+                return NoOpStreamedSpan(
+                    unsampled_reason="ignored",
+                )
+            if isinstance(parent_span, NoOpStreamedSpan):
                 return NoOpStreamedSpan(
-                    trace_id=parent_span.trace_id, unsampled_reason="ignored"
+                    unsampled_reason=parent_span._unsampled_reason,
                 )
 
             return StreamedSpan(
@@ -1302,7 +1302,7 @@ def start_streamed_span(
 
     def _update_sample_rate(self, sample_rate: float) -> None:
         # If we had to adjust the sample rate when setting the sampling decision
-        # for the spans, it needs to be updated in the propagation context too
+        # for a span, it needs to be updated in the propagation context too
         propagation_context = self.get_active_propagation_context()
         baggage = propagation_context.baggage
 

sentry_sdk/traces.py

@@ -552,10 +552,11 @@ def __init__(
         scope: "Optional[sentry_sdk.Scope]" = None,
         **kwargs: "Any",
     ) -> None:
-        self._segment = None  # type: ignore[assignment]
         self._scope = scope  # type: ignore[assignment]
         self._unsampled_reason = unsampled_reason
 
+        self._segment = None  # type: ignore[assignment]
+
         self._start()
 
     def __repr__(self) -> str:
@@ -577,6 +578,9 @@ def name(self) -> str:
     def name(self, value: str) -> None:
         pass
 
+    # XXX[span-first]: These default span_id and trace_id values will be used
+    # in outgoing requests if a noop span is active. Is that how it should be?
+
     @property
     def span_id(self) -> str:
         return "0000000000000000"