BerriAI/litellm · error · DatabricksException

Missing API Base - A call is being made to LLM Provider but

Error message

Missing API Base - A call is being made to LLM Provider but no api base is set either in the environment variables ({LLM_PROVIDER}_API_KEY) or via params

What it means

Raised when a custom Databricks endpoint is used and no api base URL could be resolved after the full chain: explicit api_base argument, DATABRICKS_API_BASE env var, and (for custom endpoints) the SDK fallback which is skipped. Note the message text incorrectly references _API_KEY; the actual missing item is the base URL.

Source

Thrown at litellm/llms/databricks/common_utils.py:364

            headers = headers or {}
            headers["Authorization"] = f"Bearer {access_token}"
            headers["Content-Type"] = "application/json"
        elif api_key is None and not headers:
            if custom_endpoint is True:
                raise DatabricksException(
                    status_code=400,
                    message="Missing API Key - A call is being made to LLM Provider but no key is set either in the environment variables ({LLM_PROVIDER}_API_KEY) or via params",
                )
            else:
                # Fallback to Databricks SDK (registers partner telemetry)
                verbose_logger.debug("Using Databricks SDK for authentication")
                api_base, headers = self._get_databricks_credentials(
                    api_base=api_base, api_key=api_key, headers=headers
                )

        if api_base is None:
            if custom_endpoint:
                raise DatabricksException(
                    status_code=400,
                    message="Missing API Base - A call is being made to LLM Provider but no api base is set either in the environment variables ({LLM_PROVIDER}_API_KEY) or via params",
                )
            else:
                api_base, headers = self._get_databricks_credentials(
                    api_base=api_base, api_key=api_key, headers=headers
                )

        if headers is None:
            headers = {
                "Authorization": f"Bearer {api_key}",
                "Content-Type": "application/json",
            }
        else:
            if api_key is not None:
                headers.update({"Authorization": f"Bearer {api_key}"})

        if api_key is not None:

View on GitHub (pinned to 6c2dcb801b)

Solutions

Set DATABRICKS_API_BASE to the full serving-endpoint URL (must include /serving-endpoints for native endpoints)
Or pass api_base in litellm_params / the completion call
For gateway URLs, confirm whether the /serving-endpoints suffix should be omitted (follow your gateway's docs)
Double-check the env var actually exports in the runtime shell/container

Example fix

# before
# custom endpoint, base never set

# after
export DATABRICKS_API_BASE="https://my-gateway.internal/databricks"
export DATABRICKS_API_KEY="gw-token"

Defensive patterns

Strategy: validation

Validate before calling

api_base = os.getenv("DATABRICKS_API_BASE") or params.get("api_base")
if custom_endpoint and not api_base:
    raise RuntimeError("Custom Databricks endpoint requires api_base (gateway URL)")

Try / catch

try:
    resp = litellm.completion(model=m, messages=msgs)
except Exception as e:
    if "Missing API Base" in str(e):
        raise ConfigError("Set DATABRICKS_API_BASE / api_base for the custom endpoint") from e
    raise

Prevention

Validate proxy config YAML at deploy time: every databricks/ model must declare api_base
Remember the message text mentions _API_KEY but the missing field is the base URL — check api_base first

When it happens

Trigger: custom_endpoint=True with no api_base argument, no DATABRICKS_API_BASE in the environment — LiteLLM will not guess a URL for a custom endpoint and raises this 400 before any request.

Common situations: Configuring Databricks-through-gateway where the gateway URL env var was forgotten; renaming env vars during a migration; YAML config indentation errors putting api_base under the wrong key.

Related errors

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/1903c0c920ff79ca. Report an issue: GitHub.