BerriAI/litellm · error · ValueError

No active custom logger found for callback name

Error message

No active custom logger found for callback name: {callback_name}

What it means

LoggingCallbackManager.get_active_custom_logger_for_callback_name() resolves a callback name to a live CustomLogger: it looks the class up in CustomLoggerRegistry, then searches active callbacks for an instance of that type. If the name maps to a class but no instance is currently attached (nothing in litellm.callbacks / the proxy callback lists), it raises ValueError. Callers include the proxy's cold-storage spend handler and standard logging payload setup.

Solutions

Instantiate and activate the logger before the request that resolves the name: litellm.callbacks.append(MyCustomLogger()) (append, don't replace)
On the proxy, declare the callback in config.yaml's callbacks list with its full module.path:ClassName so the server instantiates it at startup
Verify activation: len(litellm.logging_callback_manager.get_custom_loggers_for_type(MyCustomLogger)) > 0
Audit for code that reassigns litellm.callbacks = [...] later and wipes the active instance

Example fix

# before
# MyLogger registered in CustomLoggerRegistry, but never activated
litellm.success_callback = ["cold_storage"]  # resolving -> ValueError

# after
from my_app.callbacks import MyColdStorageLogger
litellm.callbacks.append(MyColdStorageLogger())
litellm.success_callback = ["cold_storage"]

Defensive patterns

Strategy: validation

Validate before calling

import litellm
from my_app.callbacks import MyCustomLogger

if len(litellm.logging_callback_manager.get_custom_loggers_for_type(MyCustomLogger)) == 0:
    litellm.callbacks.append(MyCustomLogger())

Try / catch

try:
    logger = litellm.logging_callback_manager.get_active_custom_logger_for_callback_name(name)
except ValueError:
    logger = MyCustomLogger()
    litellm.callbacks.append(logger)

Prevention

Register and activate together: appending the instance to litellm.callbacks right after registration
On the proxy, give every custom callback a config.yaml callbacks entry with its full module path
Never reassign litellm.callbacks wholesale after setup - append instead

When it happens

Trigger: Proxy or SDK config references a registered custom callback by name (e.g. a cold_storage custom_logger) but the class was never instantiated into the active callback lists - missing entry in config.yaml callbacks, a bad module path so the server never imported it, or litellm.callbacks reassigned/emptied after registration.

Common situations: Registering a CustomLogger class with the registry but forgetting litellm.callbacks = [MyHandler()]; proxy config listing the logger under the wrong yaml section so it never activates; tests that overwrite litellm.callbacks between requests; typos in the module path of the callback entry.

Related errors

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/ade14ae7927dc1be. Report an issue: GitHub.

Appendix: source

Thrown at litellm/litellm_core_utils/logging_callback_manager.py:478

    def get_active_custom_logger_for_callback_name(
        self,
        callback_name: _custom_logger_compatible_callbacks_literal,
    ) -> CustomLogger | None:
        """
        Get the active custom logger for a given callback name
        """
        from litellm.litellm_core_utils.custom_logger_registry import (
            CustomLoggerRegistry,
        )

        # get the custom logger class type
        custom_logger_class_type: Final = CustomLoggerRegistry.get_class_type_for_custom_logger_name(callback_name)

        # get the active custom logger
        custom_logger: Final = self.get_custom_loggers_for_type(custom_logger_class_type)

        if len(custom_logger) == 0:
            raise ValueError(f"No active custom logger found for callback name: {callback_name}")

        return custom_logger[0]

View on GitHub (pinned to 77b7c6c40c)