fix(core): add strip_null_bytes() to safe_dumps to prevent PostgreSQL 22P05 errors

Null bytes (\x00) in LLM request/response payloads cause PostgreSQL to
raise '22P05: invalid byte sequence for encoding UTF8: 0x00' when spend
logs are written to the database.

Changes:
- Add strip_null_bytes() helper to safe_json_dumps.py that recursively
  removes \x00 chars from strings, dicts, lists, tuples and sets
- Inline null byte removal into safe_dumps() _serialize() for str paths
  so all JSON serialization through safe_dumps() is automatically safe
- In spend_tracking_utils.py: replace json.dumps() with safe_dumps() for
  messages and request_body serialization; add strip_null_bytes() call
  in _sanitize_request_body_for_spend_logs_payload string handling

Centralizing the fix in safe_dumps() is more robust than ad-hoc
stripping at each call site.

Fixes #24310
Related: #21290, #15519

litellm/litellm_core_utils/safe_json_dumps.py

@@ -6,10 +6,26 @@ from pydantic import BaseModel
 from litellm.constants import DEFAULT_MAX_RECURSE_DEPTH
 
 
+def strip_null_bytes(data: Any) -> Any:
+    """Recursively remove \\x00 null bytes from strings to prevent PostgreSQL 22P05 errors."""
+    if isinstance(data, str):
+        return data.replace("\x00", "")
+    if isinstance(data, dict):
+        return {k: strip_null_bytes(v) for k, v in data.items()}
+    if isinstance(data, list):
+        return [strip_null_bytes(item) for item in data]
+    if isinstance(data, tuple):
+        return tuple(strip_null_bytes(item) for item in data)
+    if isinstance(data, set):
+        return {strip_null_bytes(item) for item in data}
+    return data
+
+
 def safe_dumps(data: Any, max_depth: int = DEFAULT_MAX_RECURSE_DEPTH) -> str:
     """
     Recursively serialize data while detecting circular references.
     If a circular reference is detected then a marker string is returned.
+    Null bytes are stripped from strings to prevent PostgreSQL 22P05 errors.
     """
 
     def _serialize(obj: Any, seen: set, depth: int) -> Any:
@@ -17,7 +33,9 @@ def safe_dumps(data: Any, max_depth: int = DEFAULT_MAX_RECURSE_DEPTH) -> str:
         if depth > max_depth:
             return "MaxDepthExceeded"
         # Base-case: if it is a primitive, simply return it.
-        if isinstance(obj, (str, int, float, bool, type(None))):
+        if isinstance(obj, str):
+            return obj.replace("\x00", "")
+        if isinstance(obj, (int, float, bool, type(None))):
             return obj
         # Check for circular reference.
         if id(obj) in seen:
@@ -51,7 +69,7 @@ def safe_dumps(data: Any, max_depth: int = DEFAULT_MAX_RECURSE_DEPTH) -> str:
         else:
             # Fall back to string conversion for non-serializable objects.
             try:
-                return str(obj)
+                return str(obj).replace("\x00", "")
             except Exception:
                 return "Unserializable Object"
 

litellm/proxy/spend_tracking/spend_tracking_utils.py

@@ -23,7 +23,7 @@ from litellm.litellm_core_utils.core_helpers import (
     get_litellm_metadata_from_kwargs,
     reconstruct_model_name,
 )
-from litellm.litellm_core_utils.safe_json_dumps import safe_dumps
+from litellm.litellm_core_utils.safe_json_dumps import safe_dumps, strip_null_bytes
 from litellm.proxy._types import SpendLogsMetadata, SpendLogsPayload
 from litellm.proxy.utils import PrismaClient, hash_token
 from litellm.types.utils import (
@@ -621,7 +621,7 @@ def _get_messages_for_spend_logs_payload(
                 messages = standard_logging_payload.get("messages")
                 if messages is not None:
                     try:
-                        return json.dumps(messages, default=str)
+                        return safe_dumps(messages)
                     except Exception:
                         return "{}"
     return "{}"
@@ -660,6 +660,7 @@ def _sanitize_request_body_for_spend_logs_payload(
         elif isinstance(value, list):
             return [_sanitize_value(item) for item in value]
         elif isinstance(value, str):
+            value = strip_null_bytes(value)
             if len(value) > max_string_length_prompt_in_db:
                 # Keep 35% from beginning and 65% from end (end is usually more important)
                 # This split ensures we keep more context from the end of conversations
@@ -804,7 +805,7 @@ def _get_proxy_server_request_for_spend_logs_payload(
                     perform_redaction(model_call_details=_request_body, result=None)
 
             _request_body = _sanitize_request_body_for_spend_logs_payload(_request_body)
-            _request_body_json_str = json.dumps(_request_body, default=str)
+            _request_body_json_str = safe_dumps(_request_body)
             if LITELLM_TRUNCATED_PAYLOAD_FIELD in _request_body_json_str:
                 verbose_proxy_logger.info(
                     "Spend Log: request body was truncated before storing in DB. %s",