Implement EventScrubber (getsentry#1943)

sl0thentr0py · antonpirker · web-flow · commit b339d838223a · 2023-03-21T10:07:51.000Z
As outlined in https://github.com/getsentry/rfcs/blob/main/text/0062-controlling-pii-and-credentials-in-sd-ks.md Co-authored-by: Anton Pirker <anton.pirker@sentry.io>
diff --git a/sentry_sdk/client.py b/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:
diff --git a/sentry_sdk/consts.py b/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
diff --git a/sentry_sdk/scrubber.py b/sentry_sdk/scrubber.py
@@ -0,0 +1,116 @@
+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",
+    "session",
+    # django
+    "csrftoken",
+    "sessionid",
+    # wsgi
+    "remote_addr",
+    "x_csrftoken",
+    "x_forwarded_for",
+    "set_cookie",
+    "cookie",
+    "authorization",
+    "x_api_key",
+    "x_forwarded_for",
+    "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)
diff --git a/sentry_sdk/serializer.py b/sentry_sdk/serializer.py
@@ -254,6 +254,8 @@ def _serialize_node_impl(
         obj, is_databag, should_repr_strings, remaining_depth, remaining_breadth
     ):
         # type: (Any, Optional[bool], Optional[bool], Optional[int], Optional[int]) -> Any
+        if isinstance(obj, AnnotatedValue):
+            should_repr_strings = False
         if should_repr_strings is None:
             should_repr_strings = _should_repr_strings()
 
diff --git a/tests/test_scrubber.py b/tests/test_scrubber.py
@@ -0,0 +1,155 @@
+import sys
+import logging
+
+from sentry_sdk import capture_exception, capture_event, start_transaction, start_span
+from sentry_sdk.utils import event_from_exception
+from sentry_sdk.scrubber import EventScrubber
+
+
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.DEBUG)
+
+
+def test_request_scrubbing(sentry_init, capture_events):
+    sentry_init()
+    events = capture_events()
+
+    try:
+        1 / 0
+    except ZeroDivisionError:
+        ev, _hint = event_from_exception(sys.exc_info())
+
+        ev["request"] = {
+            "headers": {
+                "COOKIE": "secret",
+                "authorization": "Bearer bla",
+                "ORIGIN": "google.com",
+            },
+            "cookies": {
+                "sessionid": "secret",
+                "foo": "bar",
+            },
+            "data": {
+                "token": "secret",
+                "foo": "bar",
+            },
+        }
+
+        capture_event(ev)
+
+    (event,) = events
+
+    assert event["request"] == {
+        "headers": {
+            "COOKIE": "[Filtered]",
+            "authorization": "[Filtered]",
+            "ORIGIN": "google.com",
+        },
+        "cookies": {"sessionid": "[Filtered]", "foo": "bar"},
+        "data": {"token": "[Filtered]", "foo": "bar"},
+    }
+
+    assert event["_meta"]["request"] == {
+        "headers": {
+            "COOKIE": {"": {"rem": [["!config", "s"]]}},
+            "authorization": {"": {"rem": [["!config", "s"]]}},
+        },
+        "cookies": {"sessionid": {"": {"rem": [["!config", "s"]]}}},
+        "data": {"token": {"": {"rem": [["!config", "s"]]}}},
+    }
+
+
+def test_stack_var_scrubbing(sentry_init, capture_events):
+    sentry_init()
+    events = capture_events()
+
+    try:
+        password = "supersecret"  # noqa
+        api_key = "1231231231"  # noqa
+        safe = "keepthis"  # noqa
+        1 / 0
+    except ZeroDivisionError:
+        capture_exception()
+
+    (event,) = events
+
+    frames = event["exception"]["values"][0]["stacktrace"]["frames"]
+    (frame,) = frames
+    assert frame["vars"]["password"] == "[Filtered]"
+    assert frame["vars"]["api_key"] == "[Filtered]"
+    assert frame["vars"]["safe"] == "'keepthis'"
+
+    meta = event["_meta"]["exception"]["values"]["0"]["stacktrace"]["frames"]["0"][
+        "vars"
+    ]
+    assert meta == {
+        "password": {"": {"rem": [["!config", "s"]]}},
+        "api_key": {"": {"rem": [["!config", "s"]]}},
+    }
+
+
+def test_breadcrumb_extra_scrubbing(sentry_init, capture_events):
+    sentry_init()
+    events = capture_events()
+
+    logger.info("bread", extra=dict(foo=42, password="secret"))
+    logger.critical("whoops", extra=dict(bar=69, auth="secret"))
+
+    (event,) = events
+
+    assert event["extra"]["bar"] == 69
+    assert event["extra"]["auth"] == "[Filtered]"
+
+    assert event["breadcrumbs"]["values"][0]["data"] == {
+        "foo": 42,
+        "password": "[Filtered]",
+    }
+
+    assert event["_meta"] == {
+        "extra": {"auth": {"": {"rem": [["!config", "s"]]}}},
+        "breadcrumbs": {
+            "values": {"0": {"data": {"password": {"": {"rem": [["!config", "s"]]}}}}}
+        },
+    }
+
+
+def test_span_data_scrubbing(sentry_init, capture_events):
+    sentry_init(traces_sample_rate=1.0)
+    events = capture_events()
+
+    with start_transaction(name="hi"):
+        with start_span(op="foo", description="bar") as span:
+            span.set_data("password", "secret")
+            span.set_data("datafoo", "databar")
+
+    (event,) = events
+    assert event["spans"][0]["data"] == {"password": "[Filtered]", "datafoo": "databar"}
+    assert event["_meta"] == {
+        "spans": {"0": {"data": {"password": {"": {"rem": [["!config", "s"]]}}}}}
+    }
+
+
+def test_custom_denylist(sentry_init, capture_events):
+    sentry_init(event_scrubber=EventScrubber(denylist=["my_sensitive_var"]))
+    events = capture_events()
+
+    try:
+        my_sensitive_var = "secret"  # noqa
+        safe = "keepthis"  # noqa
+        1 / 0
+    except ZeroDivisionError:
+        capture_exception()
+
+    (event,) = events
+
+    frames = event["exception"]["values"][0]["stacktrace"]["frames"]
+    (frame,) = frames
+    assert frame["vars"]["my_sensitive_var"] == "[Filtered]"
+    assert frame["vars"]["safe"] == "'keepthis'"
+
+    meta = event["_meta"]["exception"]["values"]["0"]["stacktrace"]["frames"]["0"][
+        "vars"
+    ]
+    assert meta == {
+        "my_sensitive_var": {"": {"rem": [["!config", "s"]]}},
+    }
