diff --git a/connexion/middleware/main.py b/connexion/middleware/main.py
index b5ae0c0e9..fb6284ecf 100644
--- a/connexion/middleware/main.py
+++ b/connexion/middleware/main.py
@@ -454,7 +454,9 @@ def add_error_handler(
 
         :param code_or_exception: An exception class or the status code of HTTP exceptions to
             handle.
-        :param function: Callable that will handle exception, may be async.
+        :param function: Callable that will handle exception and return a HTTP problem response.
+            An async coroutine has access to the stack traceback.  A sync function has no access
+            to the stack traceback because it is run in a threadpool.
         """
         if self.middleware_stack is not None:
             raise RuntimeError(
diff --git a/docs/exceptions.rst b/docs/exceptions.rst
index dd77f68f1..3d34f53d9 100644
--- a/docs/exceptions.rst
+++ b/docs/exceptions.rst
@@ -62,7 +62,7 @@ You can register error handlers on:
 
             .. warning::
 
-                ⚠️ **The following is not recommended as it complicates the exception handling logic,**
+                ⚠️ **The following is not recommended as it complicates the exception handling logic!**
 
             You can also register error handlers on the underlying flask application directly.
 
@@ -115,7 +115,11 @@ You can register error handlers on:
 
 .. note::
 
-    Error handlers can be ``async`` coroutines as well.
+    All the error handlers shown above are sync functions. The exception stack trace is
+    not available in a sync function because the middleware runs it in a threadpool.
+
+    Error handlers can be ``async`` coroutines as well.  Use a coroutine if the handler
+    function needs the stack trace from the exception, for example to log a traceback. 
 
 .. _Flask documentation: https://flask.palletsprojects.com/en/latest/errorhandling/#error-handlers