BerriAI/litellm · error · ValueError

Event hook {event_hook} is not in the supported event hooks

Error message

Event hook {event_hook} is not in the supported event hooks {supported_event_hooks}

What it means

The single-value branch of the same validation: when event_hook is a GuardrailEventHooks enum member (not a list and not a Mode), it must be a member of supported_event_hooks or initialization raises this ValueError. It guards against wiring a guardrail into a lifecycle stage its implementation does not cover.

Source

Thrown at litellm/integrations/custom_guardrail.py:491

            return
        if isinstance(event_hook, str):
            event_hook = GuardrailEventHooks(event_hook)
        if isinstance(event_hook, list):
            _validate_event_hook_list_is_in_supported_event_hooks(event_hook, supported_event_hooks)
        elif isinstance(event_hook, Mode):
            tag_values_flat: Final[list] = []
            for v in event_hook.tags.values():
                if isinstance(v, list):
                    tag_values_flat.extend(v)
                else:
                    tag_values_flat.append(v)
            _validate_event_hook_list_is_in_supported_event_hooks(tag_values_flat, supported_event_hooks)
            if event_hook.default:
                default_list = event_hook.default if isinstance(event_hook.default, list) else [event_hook.default]
                _validate_event_hook_list_is_in_supported_event_hooks(default_list, supported_event_hooks)
        elif isinstance(event_hook, GuardrailEventHooks):
            if event_hook not in supported_event_hooks:
                raise ValueError(f"Event hook {event_hook} is not in the supported event hooks {supported_event_hooks}")

    @staticmethod
    def _get_admin_metadata(data: dict) -> dict:
        """Return merged admin-configured key and team metadata from the request data.

        The proxy may inject admin metadata (user_api_key_metadata,
        user_api_key_team_metadata) into either ``metadata`` or
        ``litellm_metadata`` depending on endpoint. Check both so a caller
        cannot shadow admin config by pre-populating the other key.
        Key-level settings override team-level.
        """
        team_meta: dict = {}
        key_meta: dict = {}
        for key in ("metadata", "litellm_metadata"):
            # Defensive: an unparsed JSON-string metadata could leak past the
            # proxy's normal parse path; don't AttributeError on .get().
            meta = data.get(key)
            if not isinstance(meta, dict):

View on GitHub (pinned to 6c2dcb801b)

Solutions

1. Set event_hook to one of the guardrail's supported hooks (check the subclass's supported_event_hooks / docs)
2. Pass a list containing only supported hooks if you need to express intent at init time
3. Update custom guardrail subclasses to declare the hooks they truly implement before assigning them

Example fix

# before
my_guardrail = MyGuardrail(event_hook=GuardrailEventHooks.PRE_CALL)

# after
my_guardrail = MyGuardrail(event_hook=GuardrailEventHooks.DURING_CALL)

supported = guardrail.supported_event_hooks if hasattr(guardrail, 'supported_event_hooks') else []
if event_hook not in supported:
    raise ConfigError(f"use one of {[h.value for h in supported]}")

from litellm.types.guardrails import GuardrailEventHooks

def is_valid_event_hook(value: str, supported: list[GuardrailEventHooks]) -> bool:
    try:
        return GuardrailEventHooks(value) in supported
    except ValueError:
        return False

try:
    MyGuardrail(event_hook=GuardrailEventHooks.PRE_CALL)
except ValueError as e:
    if "not in the supported event hooks" in str(e):
        pick_supported_hook()
    raise

When it happens

Trigger: Constructing or configuring a guardrail with event_hook=GuardrailEventHooks.PRE_CALL when that hook is absent from the guardrail's supported_event_hooks list. Differs from 363 only in code path: direct enum value instead of a list/Mode.

Common situations: Programmatic guardrail instantiation in Python (custom guardrails registered at proxy startup) with a hardcoded event_hook copied from a different integration; version upgrades that narrowed a guardrail's supported hooks.

Related errors

• Event hook {hook} is not in the supported event hooks {suppo
• Setting tag-based guardrails is only available in litellm-en
• Invalid environment: {environment}. Please use one of the fo
• bucket_name must be provided for S3 destination
• Export format '{self.export_format}' not supported. Use 'par

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/8b5fc69955e64b0e. Report an issue: GitHub.