Repeated headers list for ASGI frameworks

CHANGELOG.md

@@ -19,6 +19,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   ([#2425](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/2425))
 - `opentelemetry-instrumentation-flask` Add `http.method` to `span.name`
   ([#2454](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/2454))
+- Record repeated HTTP headers in lists, rather than a comma separate strings for ASGI based web frameworks
+  ([#2361](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/2361))
 
 ### Added
 

instrumentation/opentelemetry-instrumentation-asgi/src/opentelemetry/instrumentation/asgi/__init__.py

@@ -129,10 +129,10 @@ def client_response_hook(span: Span, message: dict):
 
 The name of the added span attribute will follow the format ``http.request.header.<header_name>`` where ``<header_name>``
 is the normalized HTTP header name (lowercase, with ``-`` replaced by ``_``). The value of the attribute will be a
-single item list containing all the header values.
+list containing the header values.
 
 For example:
-``http.request.header.custom_request_header = ["<value1>,<value2>"]``
+``http.request.header.custom_request_header = ["<value1>", "<value2>"]``
 
 Response headers
 ****************
@@ -163,10 +163,10 @@ def client_response_hook(span: Span, message: dict):
 
 The name of the added span attribute will follow the format ``http.response.header.<header_name>`` where ``<header_name>``
 is the normalized HTTP header name (lowercase, with ``-`` replaced by ``_``). The value of the attribute will be a
-single item list containing all the header values.
+list containing the header values.
 
 For example:
-``http.response.header.custom_response_header = ["<value1>,<value2>"]``
+``http.response.header.custom_response_header = ["<value1>", "<value2>"]``
 
 Sanitizing headers
 ******************
@@ -193,9 +193,10 @@ def client_response_hook(span: Span, message: dict):
 
 import typing
 import urllib
+from collections import defaultdict
 from functools import wraps
 from timeit import default_timer
-from typing import Any, Awaitable, Callable, Tuple
+from typing import Any, Awaitable, Callable, DefaultDict, Tuple
 
 from asgiref.compatibility import guarantee_single_callable
 
@@ -339,24 +340,19 @@ def collect_custom_headers_attributes(
     sanitize: SanitizeValue,
     header_regexes: list[str],
     normalize_names: Callable[[str], str],
-) -> dict[str, str]:
+) -> dict[str, list[str]]:
     """
     Returns custom HTTP request or response headers to be added into SERVER span as span attributes.
 
     Refer specifications:
      - https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/trace/semantic_conventions/http.md#http-request-and-response-headers
     """
-    # Decode headers before processing.
-    headers: dict[str, str] = {}
+    headers: DefaultDict[str, list[str]] = defaultdict(list)
     raw_headers = scope_or_response_message.get("headers")
     if raw_headers:
-        for _key, _value in raw_headers:
-            key = _key.decode().lower()
-            value = _value.decode()
-            if key in headers:
-                headers[key] += f",{value}"
-            else:
-                headers[key] = value
+        for key, value in raw_headers:
+            # Decode headers before processing.
+            headers[key.decode()].append(value.decode())
 
     return sanitize.sanitize_header_values(
         headers,

instrumentation/opentelemetry-instrumentation-asgi/tests/test_asgi_custom_headers.py

@@ -152,7 +152,8 @@ def test_http_repeat_request_headers_in_span_attributes(self):
         span_list = self.exporter.get_finished_spans()
         expected = {
             "http.request.header.custom_test_header_1": (
-                "test-header-value-1,test-header-value-2",
+                "test-header-value-1",
+                "test-header-value-2",
             ),
         }
         span = next(span for span in span_list if span.kind == SpanKind.SERVER)
@@ -225,7 +226,8 @@ def test_http_repeat_response_headers_in_span_attributes(self):
         span_list = self.exporter.get_finished_spans()
         expected = {
             "http.response.header.custom_test_header_1": (
-                "test-header-value-1,test-header-value-2",
+                "test-header-value-1",
+                "test-header-value-2",
             ),
         }
         span = next(span for span in span_list if span.kind == SpanKind.SERVER)

instrumentation/opentelemetry-instrumentation-fastapi/src/opentelemetry/instrumentation/fastapi/__init__.py

@@ -115,7 +115,7 @@ def client_response_hook(span: Span, message: dict):
 single item list containing all the header values.
 
 For example:
-``http.request.header.custom_request_header = ["<value1>,<value2>"]``
+``http.request.header.custom_request_header = ["<value1>", "<value2>"]``
 
 Response headers
 ****************
@@ -146,10 +146,10 @@ def client_response_hook(span: Span, message: dict):
 
 The name of the added span attribute will follow the format ``http.response.header.<header_name>`` where ``<header_name>``
 is the normalized HTTP header name (lowercase, with ``-`` replaced by ``_``). The value of the attribute will be a
-single item list containing all the header values.
+list containing the header values.
 
 For example:
-``http.response.header.custom_response_header = ["<value1>,<value2>"]``
+``http.response.header.custom_response_header = ["<value1>", "<value2>"]``
 
 Sanitizing headers
 ******************

instrumentation/opentelemetry-instrumentation-fastapi/tests/test_fastapi_instrumentation.py

@@ -11,9 +11,9 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-
 import unittest
 from timeit import default_timer
+from typing import Mapping, Tuple
 from unittest.mock import patch
 
 import fastapi
@@ -549,6 +549,24 @@ def test_mark_span_internal_in_presence_of_span_from_other_framework(self):
         )
 
 
+class MultiMapping(Mapping):
+
+    def __init__(self, *items: Tuple[str, str]):
+        self._items = items
+
+    def __len__(self):
+        return len(self._items)
+
+    def __getitem__(self, __key):
+        raise NotImplementedError("use .items() instead")
+
+    def __iter__(self):
+        raise NotImplementedError("use .items() instead")
+
+    def items(self):
+        return self._items
+
+
 @patch.dict(
     "os.environ",
     {
@@ -575,13 +593,15 @@ def _create_app():
 
         @app.get("/foobar")
         async def _():
-            headers = {
-                "custom-test-header-1": "test-header-value-1",
-                "custom-test-header-2": "test-header-value-2",
-                "my-custom-regex-header-1": "my-custom-regex-value-1,my-custom-regex-value-2",
-                "My-Custom-Regex-Header-2": "my-custom-regex-value-3,my-custom-regex-value-4",
-                "My-Secret-Header": "My Secret Value",
-            }
+            headers = MultiMapping(
+                ("custom-test-header-1", "test-header-value-1"),
+                ("custom-test-header-2", "test-header-value-2"),
+                ("my-custom-regex-header-1", "my-custom-regex-value-1"),
+                ("my-custom-regex-header-1", "my-custom-regex-value-2"),
+                ("My-Custom-Regex-Header-2", "my-custom-regex-value-3"),
+                ("My-Custom-Regex-Header-2", "my-custom-regex-value-4"),
+                ("My-Secret-Header", "My Secret Value"),
+            )
             content = {"message": "hello world"}
             return JSONResponse(content=content, headers=headers)
 
@@ -657,10 +677,12 @@ def test_http_custom_response_headers_in_span_attributes(self):
                 "test-header-value-2",
             ),
             "http.response.header.my_custom_regex_header_1": (
-                "my-custom-regex-value-1,my-custom-regex-value-2",
+                "my-custom-regex-value-1",
+                "my-custom-regex-value-2",
             ),
             "http.response.header.my_custom_regex_header_2": (
-                "my-custom-regex-value-3,my-custom-regex-value-4",
+                "my-custom-regex-value-3",
+                "my-custom-regex-value-4",
             ),
             "http.response.header.my_secret_header": ("[REDACTED]",),
         }

instrumentation/opentelemetry-instrumentation-starlette/src/opentelemetry/instrumentation/starlette/__init__.py

@@ -108,10 +108,10 @@ def client_response_hook(span: Span, message: dict):
 
 The name of the added span attribute will follow the format ``http.request.header.<header_name>`` where ``<header_name>``
 is the normalized HTTP header name (lowercase, with ``-`` replaced by ``_``). The value of the attribute will be a
-single item list containing all the header values.
+list containing the header values.
 
 For example:
-``http.request.header.custom_request_header = ["<value1>,<value2>"]``
+``http.request.header.custom_request_header = ["<value1>", "<value2>"]``
 
 Response headers
 ****************
@@ -142,10 +142,10 @@ def client_response_hook(span: Span, message: dict):
 
 The name of the added span attribute will follow the format ``http.response.header.<header_name>`` where ``<header_name>``
 is the normalized HTTP header name (lowercase, with ``-`` replaced by ``_``). The value of the attribute will be a
-single item list containing all the header values.
+list containing the header values.
 
 For example:
-``http.response.header.custom_response_header = ["<value1>,<value2>"]``
+``http.response.header.custom_response_header = ["<value1>", "<value2>"]``
 
 Sanitizing headers
 ******************

util/opentelemetry-util-http/src/opentelemetry/util/http/__init__.py

@@ -18,7 +18,7 @@
 from re import IGNORECASE as RE_IGNORECASE
 from re import compile as re_compile
 from re import search
-from typing import Callable, Iterable, Optional
+from typing import Callable, Iterable, Mapping, Optional
 from urllib.parse import urlparse, urlunparse
 
 from opentelemetry.semconv.trace import SpanAttributes
@@ -87,32 +87,32 @@ def sanitize_header_value(self, header: str, value: str) -> str:
 
     def sanitize_header_values(
         self,
-        headers: dict[str, str],
+        headers: Mapping[str, str | list[str]],
         header_regexes: list[str],
         normalize_function: Callable[[str], str],
-    ) -> dict[str, str]:
-        values: dict[str, str] = {}
+    ) -> dict[str, list[str]]:
+        values: dict[str, list[str]] = {}
 
         if header_regexes:
             header_regexes_compiled = re_compile(
-                "|".join("^" + i + "$" for i in header_regexes),
+                "|".join(header_regexes),
                 RE_IGNORECASE,
             )
 
-            for header_name in list(
-                filter(
-                    header_regexes_compiled.match,
-                    headers.keys(),
-                )
-            ):
-                header_values = headers.get(header_name)
-                if header_values:
+            for header_name, header_value in headers.items():
+                if header_regexes_compiled.fullmatch(header_name):
                     key = normalize_function(header_name.lower())
-                    values[key] = [
-                        self.sanitize_header_value(
-                            header=header_name, value=header_values
-                        )
-                    ]
+                    if isinstance(header_value, str):
+                        values[key] = [
+                            self.sanitize_header_value(
+                                header_name, header_value
+                            )
+                        ]
+                    else:
+                        values[key] = [
+                            self.sanitize_header_value(header_name, value)
+                            for value in header_value
+                        ]
 
         return values