fix(dedupe): Avoid retaining strong references to user exceptions

DedupeIntegration stored the original exception instance as a fallback
when `weakref.ref(exc)` failed (e.g. for builtin exception types). That
meant a reference to every such exception lived on the context var
until the next exception replaced it, potentially pinning large
traceback frames, locals, and user objects in memory.

Instead, fall back to a `(type, hash(args))` fingerprint so dedupe can
still suppress duplicates without keeping the exception alive.

Refs PY-2381
Co-Authored-By: Claude Opus 4.7 <[email protected]>

Author: ericapisani

sentry_sdk/integrations/dedupe.py

@@ -8,11 +8,25 @@
 from typing import TYPE_CHECKING
 
 if TYPE_CHECKING:
-    from typing import Optional
+    from typing import Optional, Tuple, Type, Any
 
     from sentry_sdk._types import Event, Hint
 
 
+def _safe_args_hash(args: "Tuple[Any, ...]") -> int:
+    try:
+        return hash(args)
+    except Exception:
+        try:
+            return hash(repr(args))
+        except Exception:
+            return 0
+
+
+def _fingerprint(exc: BaseException) -> "Tuple[Type[BaseException], int]":
+    return (type(exc), _safe_args_hash(exc.args))
+
+
 class DedupeIntegration(Integration):
     identifier = "dedupe"
 
@@ -35,22 +49,23 @@ def processor(event: "Event", hint: "Optional[Hint]") -> "Optional[Event]":
                 return event
 
             last_seen = integration._last_seen.get(None)
-            if last_seen is not None:
-                # last_seen is either a weakref or the original instance
-                last_seen = (
-                    last_seen() if isinstance(last_seen, weakref.ref) else last_seen
-                )
-
             exc = exc_info[1]
-            if last_seen is exc:
+
+            is_duplicate = False
+            if isinstance(last_seen, weakref.ref):
+                is_duplicate = last_seen() is exc
+            elif isinstance(last_seen, tuple):
+                is_duplicate = last_seen == _fingerprint(exc)
+
+            if is_duplicate:
                 logger.info("DedupeIntegration dropped duplicated error event %s", exc)
                 return None
 
             # we can only weakref non builtin types
             try:
                 integration._last_seen.set(weakref.ref(exc))
             except TypeError:
-                integration._last_seen.set(exc)
+                integration._last_seen.set(_fingerprint(exc))
 
             return event