Implement EventScrubber

As outlined in https://github.com/getsentry/rfcs/blob/main/text/0062-controlling-pii-and-credentials-in-sd-ks.md

Author: sl0thentr0py

sentry_sdk/client.py

@@ -29,6 +29,7 @@
 from sentry_sdk.sessions import SessionFlusher
 from sentry_sdk.envelope import Envelope
 from sentry_sdk.profiler import has_profiling_enabled, setup_profiler
+from sentry_sdk.scrubber import EventScrubber
 
 from sentry_sdk._types import TYPE_CHECKING
 
@@ -111,6 +112,9 @@ def _get_options(*args, **kwargs):
     if rv["enable_tracing"] is True and rv["traces_sample_rate"] is None:
         rv["traces_sample_rate"] = 1.0
 
+    if rv["event_scrubber"] is None:
+        rv["event_scrubber"] = EventScrubber()
+
     return rv
 
 
@@ -249,6 +253,11 @@ def _prepare_event(
             self.options["project_root"],
         )
 
+        if event is not None:
+            event_scrubber = self.options["event_scrubber"]
+            if event_scrubber and not self.options["send_default_pii"]:
+                event_scrubber.scrub_event(event)
+
         # Postprocess the event here so that annotated types do
         # generally not surface in before_send
         if event is not None:

sentry_sdk/consts.py

@@ -133,6 +133,7 @@ def __init__(
         trace_propagation_targets=[  # noqa: B006
             MATCH_ALL
         ],  # type: Optional[Sequence[str]]
+        event_scrubber=None,  # type: Optional[sentry_sdk.scrubber.EventScrubber]
     ):
         # type: (...) -> None
         pass

sentry_sdk/scrubber.py

@@ -0,0 +1,115 @@
+from sentry_sdk.utils import (
+    capture_internal_exceptions,
+    AnnotatedValue,
+    iter_event_frames,
+)
+from sentry_sdk._compat import string_types
+from sentry_sdk._types import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from sentry_sdk._types import Event
+    from typing import Any
+    from typing import Dict
+    from typing import List
+    from typing import Optional
+
+
+DEFAULT_DENYLIST = [
+    # stolen from relay
+    "password",
+    "passwd",
+    "secret",
+    "api_key",
+    "apikey",
+    "auth",
+    "credentials",
+    "mysql_pwd",
+    "privatekey",
+    "private_key",
+    "token",
+    "ip_address",
+    # django
+    "csrftoken",
+    "sessionid",
+    # wsgi
+    "remote_addr",
+    "http_x_csrftoken",
+    "http_x_forwarded_for",
+    "http_set_cookie",
+    "http_cookie",
+    "http_authorization",
+    "http_x_api_key",
+    "http_x_forwarded_for",
+    "http_x_real_ip",
+]
+
+
+class EventScrubber(object):
+    def __init__(self, denylist=None):
+        # type: (Optional[List[str]]) -> None
+        self.denylist = DEFAULT_DENYLIST if denylist is None else denylist
+
+    def scrub_dict(self, d):
+        # type: (Dict[str, Any]) -> None
+        if not isinstance(d, dict):
+            return
+
+        for k in d.keys():
+            if isinstance(k, string_types) and k.lower() in self.denylist:
+                d[k] = AnnotatedValue.substituted_because_contains_sensitive_data()
+
+    def scrub_request(self, event):
+        # type: (Event) -> None
+        with capture_internal_exceptions():
+            if "request" in event:
+                if "headers" in event["request"]:
+                    self.scrub_dict(event["request"]["headers"])
+                if "cookies" in event["request"]:
+                    self.scrub_dict(event["request"]["cookies"])
+                if "data" in event["request"]:
+                    self.scrub_dict(event["request"]["data"])
+
+    def scrub_extra(self, event):
+        # type: (Event) -> None
+        with capture_internal_exceptions():
+            if "extra" in event:
+                self.scrub_dict(event["extra"])
+
+    def scrub_user(self, event):
+        # type: (Event) -> None
+        with capture_internal_exceptions():
+            if "user" in event:
+                self.scrub_dict(event["user"])
+
+    def scrub_breadcrumbs(self, event):
+        # type: (Event) -> None
+        with capture_internal_exceptions():
+            if "breadcrumbs" in event:
+                if "values" in event["breadcrumbs"]:
+                    for value in event["breadcrumbs"]["values"]:
+                        if "data" in value:
+                            self.scrub_dict(value["data"])
+
+    def scrub_frames(self, event):
+        # type: (Event) -> None
+        with capture_internal_exceptions():
+            for frame in iter_event_frames(event):
+                if "vars" in frame:
+                    self.scrub_dict(frame["vars"])
+
+    def scrub_spans(self, event):
+        # type: (Event) -> None
+        with capture_internal_exceptions():
+            if "spans" in event:
+                for span in event["spans"]:
+                    if "data" in span:
+                        self.scrub_dict(span["data"])
+
+    def scrub_event(self, event):
+        # type: (Event) -> None
+        self.scrub_request(event)
+        self.scrub_extra(event)
+        self.scrub_user(event)
+        self.scrub_breadcrumbs(event)
+        self.scrub_frames(event)
+        self.scrub_spans(event)