Support capture of exceptions from span event for live metrics (#39168)

Author: lzchen

sdk/monitor/azure-monitor-opentelemetry-exporter/CHANGELOG.md

@@ -10,6 +10,8 @@
   ([#38451](https://github.com/Azure/azure-sdk-for-python/pull/38451))
 - Implement live metrics filtering for docs
   ([#38925](https://github.com/Azure/azure-sdk-for-python/pull/38925))
+- Implement live metrics + filtering for span event exceptions
+  ([#39168](https://github.com/Azure/azure-sdk-for-python/pull/39168))
 
 ### Breaking Changes
 

sdk/monitor/azure-monitor-opentelemetry-exporter/azure/monitor/opentelemetry/exporter/_quickpulse/_live_metrics.py

@@ -204,7 +204,16 @@ def _record_span(self, span: ReadableSpan) -> None:
                 # Process docs for quickpulse filtering
                 _apply_document_filters_from_telemetry_data(data)
 
-                # TODO: derive exception metrics from span events
+                # Derive exception metrics from span events
+                if span.events:
+                    for event in span.events:
+                        if event.name == "exception":
+                            self._exception_rate_counter.add(1)
+                            # Derive metrics for quickpulse filtering for exception
+                            exc_data = _ExceptionData._from_span_event(event)
+                            _derive_metrics_from_telemetry_data(exc_data)
+                            # Process docs for quickpulse filtering for exception
+                            _apply_document_filters_from_telemetry_data(exc_data)
             except Exception:  # pylint: disable=broad-except
                 _logger.exception("Exception occurred while recording span.")
 

sdk/monitor/azure-monitor-opentelemetry-exporter/azure/monitor/opentelemetry/exporter/_quickpulse/_types.py

@@ -5,7 +5,7 @@
 from typing import Dict, no_type_check
 
 from opentelemetry.sdk._logs import LogRecord
-from opentelemetry.sdk.trace import ReadableSpan
+from opentelemetry.sdk.trace import Event, ReadableSpan
 from opentelemetry.semconv.trace import SpanAttributes
 from opentelemetry.trace import SpanKind
 
@@ -166,6 +166,15 @@ def _from_log_record(log_record: LogRecord):
             custom_dimensions=log_record.attributes,
         )
 
+    @staticmethod
+    @no_type_check
+    def _from_span_event(span_event: Event):
+        return _ExceptionData(
+            message=str(span_event.attributes.get(SpanAttributes.EXCEPTION_MESSAGE, "")),
+            stack_trace=str(span_event.attributes.get(SpanAttributes.EXCEPTION_STACKTRACE, "")),
+            custom_dimensions=span_event.attributes,
+        )
+
 
 @dataclass
 class _TraceData(_TelemetryData):

sdk/monitor/azure-monitor-opentelemetry-exporter/tests/quickpulse/test_live_metrics.py

@@ -352,6 +352,48 @@ def test_record_span_derive_filter_metrics(
         metric_derive_mock.assert_called_once_with(data)
         doc_mock.assert_called_once_with(data)
 
+    @mock.patch("azure.monitor.opentelemetry.exporter._quickpulse._live_metrics._ExceptionData")
+    @mock.patch("azure.monitor.opentelemetry.exporter._quickpulse._live_metrics._apply_document_filters_from_telemetry_data")
+    @mock.patch("azure.monitor.opentelemetry.exporter._quickpulse._live_metrics._derive_metrics_from_telemetry_data")
+    @mock.patch("azure.monitor.opentelemetry.exporter._quickpulse._live_metrics._TelemetryData")
+    @mock.patch("azure.monitor.opentelemetry.exporter._quickpulse._live_metrics._is_post_state")
+    def test_record_span_span_event_exception(
+        self, post_state_mock, data_mock, metric_derive_mock, doc_mock, exc_data_mock,
+    ):
+        post_state_mock.return_value = True
+        span_mock = mock.Mock()
+        event_mock = mock.Mock()
+        event_mock.name = "exception"
+        span_mock.events = [event_mock]
+        span_mock.end_time = 10
+        span_mock.start_time = 5
+        data = _RequestData(
+            custom_dimensions={},
+            duration=1000,
+            success=True,
+            name="test_req",
+            response_code=400,
+            url="test_url",
+        )
+        exc_data = _ExceptionData(
+            custom_dimensions={},
+            message="exception",
+            stack_trace="",
+        )
+        data_mock._from_span.return_value = data
+        exc_data_mock._from_span_event.return_value = exc_data
+        qpm = _QuickpulseManager(
+            connection_string="InstrumentationKey=4321abcd-5678-4efa-8abc-1234567890ac;LiveEndpoint=https://eastus.livediagnostics.monitor.azure.com/",
+            resource=Resource.create(),
+        )
+        qpm._record_span(span_mock)
+        data_mock._from_span.assert_called_once_with(span_mock)
+        exc_data_mock._from_span_event.assert_called_once_with(event_mock)
+        metric_derive_mock.assert_any_call(data)
+        doc_mock.assert_any_call(data)
+        metric_derive_mock.assert_any_call(exc_data)
+        doc_mock.assert_any_call(exc_data)
+
     @mock.patch("azure.monitor.opentelemetry.exporter._quickpulse._live_metrics._apply_document_filters_from_telemetry_data")
     @mock.patch("azure.monitor.opentelemetry.exporter._quickpulse._live_metrics._derive_metrics_from_telemetry_data")
     @mock.patch("azure.monitor.opentelemetry.exporter._quickpulse._live_metrics._TelemetryData")