Converter poc opentm instrumentation

converter/converter/converter.py

@@ -14,10 +14,14 @@
 )
 from converter.logging_config import configure_logging, LoggingKeys
 from converter.database import init_db, get_db
+from converter.tracing import configure_tracing, tag_current_span
 
 configure_logging()
 
 app = Flask(__name__)
+# Configure tracing BEFORE init_db: PymongoInstrumentor patches MongoClient.__init__, so it
+# must run before init_db() creates the client, otherwise MongoDB commands emit no spans.
+configure_tracing(app)
 init_db(app)
 
 multiproc_dir = os.getenv("PROMETHEUS_MULTIPROC_DIR")
@@ -80,6 +84,9 @@ def convert():
     except Exception:
         pass
 
+    # Label the trace span independently, so a malformed field above never prevents tagging.
+    tag_current_span(edxl_json)
+
     if not source_version or not target_version or not edxl_json:
         return raise_error(
             f"Missing required fields: sourceVersion={source_version}, targetVersion={target_version}, edxl present={edxl_json is not None}"

converter/converter/tracing.py

@@ -0,0 +1,77 @@
+import logging
+import os
+
+logger = logging.getLogger(__name__)
+
+# Span attributes, kept in sync with the dispatcher's labels so a message can be
+# followed across both services.
+CASE_ID_ATTRIBUTE = "hubsante.case_id"
+SENDER_ATTRIBUTE = "hubsante.sender"
+RECIPIENT_ATTRIBUTE = "hubsante.recipient"
+USE_CASE_ATTRIBUTE = "hubsante.use_case"
+
+
+def is_tracing_enabled() -> bool:
+    return os.getenv("OTEL_SDK_DISABLED", "false").strip().lower() != "true"
+
+
+def tag_current_span(edxl_json) -> None:
+    from opentelemetry import trace
+
+    from converter.utils import (
+        extract_case_id,
+        extract_message_content,
+        extract_message_type_from_message_content,
+        get_recipient,
+        get_sender,
+    )
+
+    span = trace.get_current_span()
+
+    def use_case():
+        return extract_message_type_from_message_content(
+            extract_message_content(edxl_json)
+        )
+
+    for attribute, getter in (
+        (SENDER_ATTRIBUTE, lambda: get_sender(edxl_json)),
+        (RECIPIENT_ATTRIBUTE, lambda: get_recipient(edxl_json)),
+        (USE_CASE_ATTRIBUTE, use_case),
+        (CASE_ID_ATTRIBUTE, lambda: extract_case_id(edxl_json)),
+    ):
+        try:
+            value = getter()
+        except Exception:
+            value = None
+        if value:
+            span.set_attribute(attribute, value)
+
+
+def configure_tracing(app) -> None:
+    if not is_tracing_enabled():
+        logger.info("OpenTelemetry tracing disabled via OTEL_SDK_DISABLED")
+        return
+
+    from opentelemetry import trace
+    from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
+    from opentelemetry.instrumentation.flask import FlaskInstrumentor
+    from opentelemetry.instrumentation.pymongo import PymongoInstrumentor
+    from opentelemetry.sdk.resources import SERVICE_NAME, SERVICE_VERSION, Resource
+    from opentelemetry.sdk.trace import TracerProvider
+    from opentelemetry.sdk.trace.export import BatchSpanProcessor
+
+    resource = Resource.create(
+        {
+            SERVICE_NAME: os.getenv("OTEL_SERVICE_NAME", "converter"),
+            SERVICE_VERSION: os.getenv("CONVERTER_VERSION", "unknown"),
+        }
+    )
+    provider = TracerProvider(resource=resource)
+    provider.add_span_processor(BatchSpanProcessor(OTLPSpanExporter()))
+    trace.set_tracer_provider(provider)
+
+    FlaskInstrumentor().instrument_app(app)
+    PymongoInstrumentor().instrument()
+    logger.info(
+        f"OpenTelemetry tracing configured for service '{resource.attributes.get(SERVICE_NAME)}'"
+    )

converter/converter/utils.py

@@ -47,6 +47,21 @@ def extract_message_type_from_message_content(
     return filtered_keys[0]
 
 
+def extract_case_id(edxl_json) -> Optional[str]:
+    """Best-effort extraction of the business caseId, used as a span label.
+
+    The caseId lives directly under the use-case object (e.g. ``resourcesInfo.caseId``).
+    Returns ``None`` when the message carries no caseId.
+    """
+    try:
+        message_content = extract_message_content(edxl_json)
+        message_type = extract_message_type_from_message_content(message_content)
+        case_id = message_content.get(message_type, {}).get("caseId")
+        return case_id if isinstance(case_id, str) and case_id else None
+    except Exception:
+        return None
+
+
 def delete_paths(data: Dict[str, Any], paths: List[str]) -> None:
     """
     Safely deletes keys in a dictionary based on dot-separated paths.

converter/pyproject.toml

@@ -15,6 +15,10 @@ dependencies = [
     "pyyaml>=6.0.2",
     "typing-extensions>=4.15.0",
     "werkzeug>=3.1.3",
+    "opentelemetry-sdk>=1.43.0",
+    "opentelemetry-exporter-otlp-proto-http>=1.43.0",
+    "opentelemetry-instrumentation-flask>=0.64b0",
+    "opentelemetry-instrumentation-pymongo>=0.64b0",
 ]
 
 [tool.pytest.ini_options]

converter/tests/conftest.py

@@ -1,5 +1,10 @@
+import os
 from unittest.mock import MagicMock, patch
 
+# Disable OpenTelemetry tracing during tests (no exporter connections, no Flask
+# instrumentation). Must be set before the app module is imported below.
+os.environ.setdefault("OTEL_SDK_DISABLED", "true")
+
 # Patch init_db before the app module is imported, so it doesn't attempt
 # to connect to MongoDB during tests.
 with patch("converter.database.init_db") as mock_init_db:

converter/tests/test_utils.py

@@ -10,6 +10,7 @@
     translate_key_words,
     update_json_value,
     set_value,
+    extract_case_id,
     extract_message_type_from_message_content,
     switch_field_name,
 )
@@ -71,6 +72,29 @@ def test_format_object_nested():
     assert result == expected
 
 
+def _edxl_with_message(message):
+    return {"content": [{"jsonContent": {"embeddedJsonContent": {"message": message}}}]}
+
+
+def test_extract_case_id_present():
+    edxl = _edxl_with_message({"resourcesInfo": {"caseId": "case-123"}})
+    assert extract_case_id(edxl) == "case-123"
+
+
+def test_extract_case_id_absent():
+    edxl = _edxl_with_message({"technicalNoreq": {"optionalStringField": "value"}})
+    assert extract_case_id(edxl) is None
+
+
+def test_extract_case_id_empty_string():
+    edxl = _edxl_with_message({"resourcesInfo": {"caseId": ""}})
+    assert extract_case_id(edxl) is None
+
+
+def test_extract_case_id_malformed_payload():
+    assert extract_case_id({}) is None
+
+
 class TestDeletePaths(unittest.TestCase):
     def test_delete_paths(self):
         data = {"a": {"b": {"c": 1, "d": 2}, "e": 3}, "f": 4}