sentry-sdk 0.7.5__py2.py3-none-any.whl → 2.46.0__py2.py3-none-any.whl

sentry_sdk/integrations/wsgi.py

@@ -1,165 +1,245 @@
 import sys
-
-from sentry_sdk.hub import Hub, _should_send_default_pii
-from sentry_sdk.utils import capture_internal_exceptions, event_from_exception
-from sentry_sdk._compat import PY2, reraise
-from sentry_sdk.integrations._wsgi_common import _filter_headers
-
-if False:
+from functools import partial
+
+import sentry_sdk
+from sentry_sdk._werkzeug import get_host, _get_headers
+from sentry_sdk.api import continue_trace
+from sentry_sdk.consts import OP
+from sentry_sdk.scope import should_send_default_pii
+from sentry_sdk.integrations._wsgi_common import (
+    DEFAULT_HTTP_METHODS_TO_CAPTURE,
+    _filter_headers,
+    nullcontext,
+)
+from sentry_sdk.sessions import track_session
+from sentry_sdk.scope import use_isolation_scope
+from sentry_sdk.tracing import Transaction, TransactionSource
+from sentry_sdk.utils import (
+    ContextVar,
+    capture_internal_exceptions,
+    event_from_exception,
+    reraise,
+)
+
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
     from typing import Callable
     from typing import Dict
-    from typing import List
     from typing import Iterator
     from typing import Any
     from typing import Tuple
     from typing import Optional
+    from typing import TypeVar
+    from typing import Protocol
 
     from sentry_sdk.utils import ExcInfo
+    from sentry_sdk._types import Event, EventProcessor
 
+    WsgiResponseIter = TypeVar("WsgiResponseIter")
+    WsgiResponseHeaders = TypeVar("WsgiResponseHeaders")
+    WsgiExcInfo = TypeVar("WsgiExcInfo")
 
-if PY2:
-
-    def wsgi_decoding_dance(s, charset="utf-8", errors="replace"):
-        # type: (str, str, str) -> str
-        return s.decode(charset, errors)
-
+    class StartResponse(Protocol):
+        def __call__(self, status, response_headers, exc_info=None):  # type: ignore
+            # type: (str, WsgiResponseHeaders, Optional[WsgiExcInfo]) -> WsgiResponseIter
+            pass
 
-else:
 
-    def wsgi_decoding_dance(s, charset="utf-8", errors="replace"):
-        # type: (str, str, str) -> str
-        return s.encode("latin1").decode(charset, errors)
+_wsgi_middleware_applied = ContextVar("sentry_wsgi_middleware_applied")
 
 
-def get_host(environ):
-    # type: (Dict[str, str]) -> str
-    """Return the host for the given WSGI environment. Yanked from Werkzeug."""
-    if environ.get("HTTP_HOST"):
-        rv = environ["HTTP_HOST"]
-        if environ["wsgi.url_scheme"] == "http" and rv.endswith(":80"):
-            rv = rv[:-3]
-        elif environ["wsgi.url_scheme"] == "https" and rv.endswith(":443"):
-            rv = rv[:-4]
-    elif environ.get("SERVER_NAME"):
-        rv = environ["SERVER_NAME"]
-        if (environ["wsgi.url_scheme"], environ["SERVER_PORT"]) not in (
-            ("https", "443"),
-            ("http", "80"),
-        ):
-            rv += ":" + environ["SERVER_PORT"]
-    else:
-        # In spite of the WSGI spec, SERVER_NAME might not be present.
-        rv = "unknown"
-
-    return rv
+def wsgi_decoding_dance(s, charset="utf-8", errors="replace"):
+    # type: (str, str, str) -> str
+    return s.encode("latin1").decode(charset, errors)
 
 
-def get_request_url(environ):
-    # type: (Dict[str, str]) -> str
+def get_request_url(environ, use_x_forwarded_for=False):
+    # type: (Dict[str, str], bool) -> str
     """Return the absolute URL without query string for the given WSGI
     environment."""
+    script_name = environ.get("SCRIPT_NAME", "").rstrip("/")
+    path_info = environ.get("PATH_INFO", "").lstrip("/")
+    path = f"{script_name}/{path_info}"
+
     return "%s://%s/%s" % (
         environ.get("wsgi.url_scheme"),
-        get_host(environ),
-        wsgi_decoding_dance(environ.get("PATH_INFO") or "").lstrip("/"),
+        get_host(environ, use_x_forwarded_for),
+        wsgi_decoding_dance(path).lstrip("/"),
     )
 
 
-class SentryWsgiMiddleware(object):
-    __slots__ = ("app",)
+class SentryWsgiMiddleware:
+    __slots__ = (
+        "app",
+        "use_x_forwarded_for",
+        "span_origin",
+        "http_methods_to_capture",
+    )
 
-    def __init__(self, app):
-        # type: (Callable) -> None
+    def __init__(
+        self,
+        app,  # type: Callable[[Dict[str, str], Callable[..., Any]], Any]
+        use_x_forwarded_for=False,  # type: bool
+        span_origin="manual",  # type: str
+        http_methods_to_capture=DEFAULT_HTTP_METHODS_TO_CAPTURE,  # type: Tuple[str, ...]
+    ):
+        # type: (...) -> None
         self.app = app
+        self.use_x_forwarded_for = use_x_forwarded_for
+        self.span_origin = span_origin
+        self.http_methods_to_capture = http_methods_to_capture
 
     def __call__(self, environ, start_response):
-        # type: (Dict[str, str], Callable) -> _ScopedResponse
-        hub = Hub(Hub.current)
-
-        with hub:
-            with capture_internal_exceptions():
-                with hub.configure_scope() as scope:
-                    scope._name = "wsgi"
-                    scope.add_event_processor(_make_wsgi_event_processor(environ))
-
-            try:
-                rv = self.app(environ, start_response)
-            except Exception:
-                reraise(*_capture_exception(hub))
-
-        return _ScopedResponse(hub, rv)
+        # type: (Dict[str, str], Callable[..., Any]) -> _ScopedResponse
+        if _wsgi_middleware_applied.get(False):
+            return self.app(environ, start_response)
+
+        _wsgi_middleware_applied.set(True)
+        try:
+            with sentry_sdk.isolation_scope() as scope:
+                with track_session(scope, session_mode="request"):
+                    with capture_internal_exceptions():
+                        scope.clear_breadcrumbs()
+                        scope._name = "wsgi"
+                        scope.add_event_processor(
+                            _make_wsgi_event_processor(
+                                environ, self.use_x_forwarded_for
+                            )
+                        )
+
+                    method = environ.get("REQUEST_METHOD", "").upper()
+                    transaction = None
+                    if method in self.http_methods_to_capture:
+                        transaction = continue_trace(
+                            environ,
+                            op=OP.HTTP_SERVER,
+                            name="generic WSGI request",
+                            source=TransactionSource.ROUTE,
+                            origin=self.span_origin,
+                        )
+
+                    transaction_context = (
+                        sentry_sdk.start_transaction(
+                            transaction,
+                            custom_sampling_context={"wsgi_environ": environ},
+                        )
+                        if transaction is not None
+                        else nullcontext()
+                    )
+                    with transaction_context:
+                        try:
+                            response = self.app(
+                                environ,
+                                partial(
+                                    _sentry_start_response, start_response, transaction
+                                ),
+                            )
+                        except BaseException:
+                            reraise(*_capture_exception())
+        finally:
+            _wsgi_middleware_applied.set(False)
+
+        return _ScopedResponse(scope, response)
+
+
+def _sentry_start_response(  # type: ignore
+    old_start_response,  # type: StartResponse
+    transaction,  # type: Optional[Transaction]
+    status,  # type: str
+    response_headers,  # type: WsgiResponseHeaders
+    exc_info=None,  # type: Optional[WsgiExcInfo]
+):
+    # type: (...) -> WsgiResponseIter
+    with capture_internal_exceptions():
+        status_int = int(status.split(" ", 1)[0])
+        if transaction is not None:
+            transaction.set_http_status(status_int)
+
+    if exc_info is None:
+        # The Django Rest Framework WSGI test client, and likely other
+        # (incorrect) implementations, cannot deal with the exc_info argument
+        # if one is present. Avoid providing a third argument if not necessary.
+        return old_start_response(status, response_headers)
+    else:
+        return old_start_response(status, response_headers, exc_info)
 
 
 def _get_environ(environ):
     # type: (Dict[str, str]) -> Iterator[Tuple[str, str]]
     """
-    Returns our whitelisted environment variables.
+    Returns our explicitly included environment variables we want to
+    capture (server name, port and remote addr if pii is enabled).
     """
-    keys = ("SERVER_NAME", "SERVER_PORT")
-    if _should_send_default_pii():
-        keys += ("REMOTE_ADDR",)  # type: ignore
+    keys = ["SERVER_NAME", "SERVER_PORT"]
+    if should_send_default_pii():
+        # make debugging of proxy setup easier. Proxy headers are
+        # in headers.
+        keys += ["REMOTE_ADDR"]
 
     for key in keys:
         if key in environ:
             yield key, environ[key]
 
 
-# `get_headers` comes from `werkzeug.datastructures.EnvironHeaders`
-#
-# We need this function because Django does not give us a "pure" http header
-# dict. So we might as well use it for all WSGI integrations.
-def _get_headers(environ):
-    # type: (Dict[str, str]) -> Iterator[Tuple[str, str]]
-    """
-    Returns only proper HTTP headers.
-
-    """
-    for key, value in environ.items():
-        key = str(key)
-        if key.startswith("HTTP_") and key not in (
-            "HTTP_CONTENT_TYPE",
-            "HTTP_CONTENT_LENGTH",
-        ):
-            yield key[5:].replace("_", "-").title(), value
-        elif key in ("CONTENT_TYPE", "CONTENT_LENGTH"):
-            yield key.replace("_", "-").title(), value
-
-
 def get_client_ip(environ):
     # type: (Dict[str, str]) -> Optional[Any]
     """
-    Naively yank the first IP address in an X-Forwarded-For header
-    and assume this is correct.
-
-    Note: Don't use this in security sensitive situations since this
-    value may be forged from a client.
+    Infer the user IP address from various headers. This cannot be used in
+    security sensitive situations since the value may be forged from a client,
+    but it's good enough for the event payload.
     """
     try:
         return environ["HTTP_X_FORWARDED_FOR"].split(",")[0].strip()
     except (KeyError, IndexError):
-        return environ.get("REMOTE_ADDR")
+        pass
 
+    try:
+        return environ["HTTP_X_REAL_IP"]
+    except KeyError:
+        pass
+
+    return environ.get("REMOTE_ADDR")
 
-def _capture_exception(hub):
-    # type: (Hub) -> ExcInfo
-    # Check client here as it might have been unset while streaming response
-    if hub.client is not None:
-        exc_info = sys.exc_info()
+
+def _capture_exception():
+    # type: () -> ExcInfo
+    """
+    Captures the current exception and sends it to Sentry.
+    Returns the ExcInfo tuple to it can be reraised afterwards.
+    """
+    exc_info = sys.exc_info()
+    e = exc_info[1]
+
+    # SystemExit(0) is the only uncaught exception that is expected behavior
+    should_skip_capture = isinstance(e, SystemExit) and e.code in (0, None)
+    if not should_skip_capture:
         event, hint = event_from_exception(
             exc_info,
-            client_options=hub.client.options,
+            client_options=sentry_sdk.get_client().options,
             mechanism={"type": "wsgi", "handled": False},
         )
-        hub.capture_event(event, hint=hint)
+        sentry_sdk.capture_event(event, hint=hint)
+
     return exc_info
 
 
-class _ScopedResponse(object):
-    __slots__ = ("_response", "_hub")
+class _ScopedResponse:
+    """
+    Users a separate scope for each response chunk.
+
+    This will make WSGI apps more tolerant against:
+    - WSGI servers streaming responses from a different thread/from
+      different threads than the one that called start_response
+    - close() not being called
+    - WSGI servers streaming responses interleaved from the same thread
+    """
+
+    __slots__ = ("_response", "_scope")
 
-    def __init__(self, hub, response):
-        # type: (Hub, List[bytes]) -> None
-        self._hub = hub
+    def __init__(self, scope, response):
+        # type: (sentry_sdk.scope.Scope, Iterator[bytes]) -> None
+        self._scope = scope
         self._response = response
 
     def __iter__(self):
@@ -167,32 +247,33 @@ class _ScopedResponse(object):
         iterator = iter(self._response)
 
         while True:
-            with self._hub:
+            with use_isolation_scope(self._scope):
                 try:
                     chunk = next(iterator)
                 except StopIteration:
                     break
-                except Exception:
-                    reraise(*_capture_exception(self._hub))
+                except BaseException:
+                    reraise(*_capture_exception())
 
             yield chunk
 
     def close(self):
-        with self._hub:
+        # type: () -> None
+        with use_isolation_scope(self._scope):
             try:
-                self._response.close()
+                self._response.close()  # type: ignore
             except AttributeError:
                 pass
-            except Exception:
-                reraise(*_capture_exception(self._hub))
+            except BaseException:
+                reraise(*_capture_exception())
 
 
-def _make_wsgi_event_processor(environ):
-    # type: (Dict[str, str]) -> Callable
+def _make_wsgi_event_processor(environ, use_x_forwarded_for):
+    # type: (Dict[str, str], bool) -> EventProcessor
     # It's a bit unfortunate that we have to extract and parse the request data
     # from the environ so eagerly, but there are a few good reasons for this.
     #
-    # We might be in a situation where the scope/hub never gets torn down
+    # We might be in a situation where the scope never gets torn down
     # properly. In that case we will have an unnecessary strong reference to
     # all objects in the environ (some of which may take a lot of memory) when
     # we're really just interested in a few of them.
@@ -202,21 +283,22 @@ def _make_wsgi_event_processor(environ):
     # https://github.com/unbit/uwsgi/issues/1950
 
     client_ip = get_client_ip(environ)
-    request_url = get_request_url(environ)
+    request_url = get_request_url(environ, use_x_forwarded_for)
     query_string = environ.get("QUERY_STRING")
     method = environ.get("REQUEST_METHOD")
     env = dict(_get_environ(environ))
     headers = _filter_headers(dict(_get_headers(environ)))
 
     def event_processor(event, hint):
-        # type: (Dict[str, Any], Dict[str, Any]) -> Dict[str, Any]
+        # type: (Event, Dict[str, Any]) -> Event
         with capture_internal_exceptions():
             # if the code below fails halfway through we at least have some data
             request_info = event.setdefault("request", {})
 
-            if _should_send_default_pii():
+            if should_send_default_pii():
                 user_info = event.setdefault("user", {})
-                user_info["ip_address"] = client_ip
+                if client_ip:
+                    user_info.setdefault("ip_address", client_ip)
 
             request_info["url"] = request_url
             request_info["query_string"] = query_string

sentry_sdk/logger.py

@@ -0,0 +1,96 @@
+# NOTE: this is the logger sentry exposes to users, not some generic logger.
+import functools
+import time
+from typing import Any
+
+from sentry_sdk import get_client
+from sentry_sdk.utils import safe_repr, capture_internal_exceptions
+
+OTEL_RANGES = [
+    # ((severity level range), severity text)
+    # https://opentelemetry.io/docs/specs/otel/logs/data-model
+    ((1, 4), "trace"),
+    ((5, 8), "debug"),
+    ((9, 12), "info"),
+    ((13, 16), "warn"),
+    ((17, 20), "error"),
+    ((21, 24), "fatal"),
+]
+
+
+class _dict_default_key(dict):  # type: ignore[type-arg]
+    """dict that returns the key if missing."""
+
+    def __missing__(self, key):
+        # type: (str) -> str
+        return "{" + key + "}"
+
+
+def _capture_log(severity_text, severity_number, template, **kwargs):
+    # type: (str, int, str, **Any) -> None
+    client = get_client()
+
+    body = template
+    attrs = {}  # type: dict[str, str | bool | float | int]
+    if "attributes" in kwargs:
+        attrs.update(kwargs.pop("attributes"))
+    for k, v in kwargs.items():
+        attrs[f"sentry.message.parameter.{k}"] = v
+    if kwargs:
+        # only attach template if there are parameters
+        attrs["sentry.message.template"] = template
+
+        with capture_internal_exceptions():
+            body = template.format_map(_dict_default_key(kwargs))
+
+    attrs = {
+        k: (
+            v
+            if (
+                isinstance(v, str)
+                or isinstance(v, int)
+                or isinstance(v, bool)
+                or isinstance(v, float)
+            )
+            else safe_repr(v)
+        )
+        for (k, v) in attrs.items()
+    }
+
+    # noinspection PyProtectedMember
+    client._capture_log(
+        {
+            "severity_text": severity_text,
+            "severity_number": severity_number,
+            "attributes": attrs,
+            "body": body,
+            "time_unix_nano": time.time_ns(),
+            "trace_id": None,
+        },
+    )
+
+
+trace = functools.partial(_capture_log, "trace", 1)
+debug = functools.partial(_capture_log, "debug", 5)
+info = functools.partial(_capture_log, "info", 9)
+warning = functools.partial(_capture_log, "warn", 13)
+error = functools.partial(_capture_log, "error", 17)
+fatal = functools.partial(_capture_log, "fatal", 21)
+
+
+def _otel_severity_text(otel_severity_number):
+    # type: (int) -> str
+    for (lower, upper), severity in OTEL_RANGES:
+        if lower <= otel_severity_number <= upper:
+            return severity
+
+    return "default"
+
+
+def _log_level_to_otel(level, mapping):
+    # type: (int, dict[Any, int]) -> tuple[int, str]
+    for py_level, otel_severity_number in sorted(mapping.items(), reverse=True):
+        if level >= py_level:
+            return otel_severity_number, _otel_severity_text(otel_severity_number)
+
+    return 0, "default"

sentry_sdk/metrics.py

@@ -0,0 +1,81 @@
+"""
+NOTE: This file contains experimental code that may be changed or removed at any
+time without prior notice.
+"""
+
+import time
+from typing import Any, Optional, TYPE_CHECKING, Union
+
+import sentry_sdk
+from sentry_sdk.utils import safe_repr
+
+if TYPE_CHECKING:
+    from sentry_sdk._types import Metric, MetricType
+
+
+def _capture_metric(
+    name,  # type: str
+    metric_type,  # type: MetricType
+    value,  # type: float
+    unit=None,  # type: Optional[str]
+    attributes=None,  # type: Optional[dict[str, Any]]
+):
+    # type: (...) -> None
+    client = sentry_sdk.get_client()
+
+    attrs = {}  # type: dict[str, Union[str, bool, float, int]]
+    if attributes:
+        for k, v in attributes.items():
+            attrs[k] = (
+                v
+                if (
+                    isinstance(v, str)
+                    or isinstance(v, int)
+                    or isinstance(v, bool)
+                    or isinstance(v, float)
+                )
+                else safe_repr(v)
+            )
+
+    metric = {
+        "timestamp": time.time(),
+        "trace_id": None,
+        "span_id": None,
+        "name": name,
+        "type": metric_type,
+        "value": float(value),
+        "unit": unit,
+        "attributes": attrs,
+    }  # type: Metric
+
+    client._capture_metric(metric)
+
+
+def count(
+    name,  # type: str
+    value,  # type: float
+    unit=None,  # type: Optional[str]
+    attributes=None,  # type: Optional[dict[str, Any]]
+):
+    # type: (...) -> None
+    _capture_metric(name, "counter", value, unit, attributes)
+
+
+def gauge(
+    name,  # type: str
+    value,  # type: float
+    unit=None,  # type: Optional[str]
+    attributes=None,  # type: Optional[dict[str, Any]]
+):
+    # type: (...) -> None
+    _capture_metric(name, "gauge", value, unit, attributes)
+
+
+def distribution(
+    name,  # type: str
+    value,  # type: float
+    unit=None,  # type: Optional[str]
+    attributes=None,  # type: Optional[dict[str, Any]]
+):
+    # type: (...) -> None
+    _capture_metric(name, "distribution", value, unit, attributes)