Adding trace identification numbers (#888)

* Adding trace identification numbers to select log statements to allow for easier categorization in log handlers

* Reformat

* Fix import formatting

* Remove key and enum definitions

* Fix copy paste bug

* Switch workflow test to logcapturer

* Remove test artifact

* Fixing test

* Update temporalio/types.py

Co-authored-by: Chad Retz <chad@temporal.io>

---------

Co-authored-by: Chad Retz <chad@temporal.io>

Author: tconley1428

temporalio/worker/_activity.py

@@ -351,11 +351,15 @@ async def _run_activity(
                     ):
                         # Downgrade log level to DEBUG for BENIGN application errors.
                         temporalio.activity.logger.debug(
-                            "Completing activity as failed", exc_info=True
+                            "Completing activity as failed",
+                            exc_info=True,
+                            extra={"__temporal_error_identifier": "ActivityFailure"},
                         )
                     else:
                         temporalio.activity.logger.warning(
-                            "Completing activity as failed", exc_info=True
+                            "Completing activity as failed",
+                            exc_info=True,
+                            extra={"__temporal_error_identifier": "ActivityFailure"},
                         )
                     await self._data_converter.encode_failure(
                         err, completion.result.failed.failure

temporalio/worker/_workflow_instance.py

@@ -447,7 +447,10 @@ def activate(
             logger.warning(
                 f"Failed activation on workflow {self._info.workflow_type} with ID {self._info.workflow_id} and run ID {self._info.run_id}",
                 exc_info=activation_err,
-                extra={"temporal_workflow": self._info._logger_details()},
+                extra={
+                    "temporal_workflow": self._info._logger_details(),
+                    "__temporal_error_identifier": "WorkflowTaskFailure",
+                },
             )
             # Set completion failure
             self._current_completion.failed.failure.SetInParent()

tests/worker/test_activity.py

@@ -1379,3 +1379,40 @@ def assert_activity_application_error(
     ret = assert_activity_error(err)
     assert isinstance(ret, ApplicationError)
     return ret
+
+
+class CustomLogHandler(logging.Handler):
+    def __init__(self):
+        super().__init__()
+        self._trace_identifiers = 0
+
+    def emit(self, record: logging.LogRecord) -> None:
+        if (
+            hasattr(record, "__temporal_error_identifier")
+            and getattr(record, "__temporal_error_identifier") == "ActivityFailure"
+        ):
+            assert record.msg.startswith("Completing activity as failed")
+            self._trace_identifiers += 1
+        return None
+
+
+async def test_activity_failure_trace_identifier(
+    client: Client, worker: ExternalWorker
+):
+    @activity.defn
+    async def raise_error():
+        raise RuntimeError("oh no!")
+
+    handler = CustomLogHandler()
+    activity.logger.base_logger.addHandler(handler)
+
+    try:
+        with pytest.raises(WorkflowFailureError) as err:
+            await _execute_workflow_with_activity(client, worker, raise_error)
+        assert (
+            str(assert_activity_application_error(err.value)) == "RuntimeError: oh no!"
+        )
+        assert handler._trace_identifiers == 1
+
+    finally:
+        activity.logger.base_logger.removeHandler(CustomLogHandler())

tests/worker/test_workflow.py

@@ -1961,8 +1961,13 @@ def logs_captured(self, *loggers: logging.Logger):
                 l.setLevel(prev_levels[i])
 
     def find_log(self, starts_with: str) -> Optional[logging.LogRecord]:
+        return self.find(lambda l: l.message.startswith(starts_with))
+
+    def find(
+        self, pred: Callable[[logging.LogRecord], bool]
+    ) -> Optional[logging.LogRecord]:
         for record in cast(List[logging.LogRecord], self.log_queue.queue):
-            if record.message.startswith(starts_with):
+            if pred(record):
                 return record
         return None
 
@@ -2058,6 +2063,7 @@ async def run(self) -> None:
         if not task_fail_once_workflow_has_failed:
             task_fail_once_workflow_has_failed = True
             raise RuntimeError("Intentional workflow task failure")
+        task_fail_once_workflow_has_failed = False
 
         # Execute activity that will fail once
         await workflow.execute_activity(
@@ -7975,6 +7981,33 @@ async def test_quick_activity_swallows_cancellation(client: Client):
         temporalio.worker._workflow_instance._raise_on_cancelling_completed_activity_override = False
 
 
+async def test_workflow_logging_trace_identifier(client: Client):
+    with LogCapturer().logs_captured(
+        temporalio.worker._workflow_instance.logger
+    ) as capturer:
+        async with new_worker(
+            client,
+            TaskFailOnceWorkflow,
+            activities=[task_fail_once_activity],
+        ) as worker:
+            await client.execute_workflow(
+                TaskFailOnceWorkflow.run,
+                id=f"workflow_failure_trace_identifier",
+                task_queue=worker.task_queue,
+            )
+
+        def workflow_failure(l: logging.LogRecord):
+            if (
+                hasattr(l, "__temporal_error_identifier")
+                and getattr(l, "__temporal_error_identifier") == "WorkflowTaskFailure"
+            ):
+                assert l.msg.startswith("Failed activation on workflow")
+                return True
+            return False
+
+        assert capturer.find(workflow_failure) is not None
+
+
 @activity.defn
 def use_in_workflow() -> bool:
     return workflow.in_workflow()