SpanProcessor.on_end exceptions escape cleanup and replace application exceptions #5624

Steps to Reproduce

from opentelemetry.sdk.trace import SpanProcessor, TracerProvider

class RaisingProcessor(SpanProcessor):
    def on_end(self, span):
        raise RuntimeError("processor failed")

provider = TracerProvider()
provider.add_span_processor(RaisingProcessor())
tracer = provider.get_tracer("reproduction")

with tracer.start_as_current_span("operation"):
    raise ValueError("application failed")

Expected Result

OnEnd(Span):

This method MUST be called synchronously within the Span.End() API, therefore it should not block or throw an exception.

Error handling:

OpenTelemetry implementations MUST NOT throw unhandled exceptions at runtime.

API methods that accept external callbacks MUST handle all errors.

SynchronousMultiSpanProcessor.on_end invokes processors without an exception guard. The exception escapes through Span.end during context-manager cleanup.

Actual Result

The caller receives RuntimeError("processor failed"), with the application’s ValueError in __context__. Replacing the application’s raise with pass also propagates RuntimeError.

Would you like to implement a fix?

--- a/opentelemetry-sdk/src/opentelemetry/sdk/trace/__init__.py
+++ b/opentelemetry-sdk/src/opentelemetry/sdk/trace/__init__.py
@@ -187,3 +187,6 @@
     def on_end(self, span: "ReadableSpan") -> None:
         for sp in self._span_processors:
-            sp.on_end(span)
+            try:
+                sp.on_end(span)
+            except Exception:
+                logger.exception("Exception while calling SpanProcessor.on_end.")

Tip

_{React with 👍 to help prioritize this issue. Please use comments to provide useful context, avoiding +1 or me too, to help us triage it. Learn more here.}