BerriAI/litellm · error · ValueError

Could not resolve credentials

Error message

Could not resolve credentials

What it means

Generic sentinel when the async load-and-cache credential path fails: load_auth threw (via asyncify) and no more specific error applies, so no (credentials, project_id) tuple can be produced or cached for the request.

Source

Thrown at litellm/llms/vertex_ai/vertex_llm_base.py:488

        return _TokenState.INVALID

    async def _load_and_cache_credentials(
        self,
        credentials: VERTEX_CREDENTIALS_TYPES | None,
        project_id: str | None,
        credential_cache_key: tuple,
    ) -> tuple[Any, str | None]:
        """Load credentials via load_auth (in thread) and cache the result."""
        try:
            _credentials, credential_project_id = await asyncify(self.load_auth)(
                credentials=credentials,
                project_id=project_id,
            )
        except Exception as e:
            verbose_logger.exception("Failed to load vertex credentials: %s", str(e))
            raise
        if _credentials is None:
            raise ValueError("Could not resolve credentials")
        self._credentials_project_mapping[credential_cache_key] = (
            _credentials,
            credential_project_id,
        )
        return _credentials, credential_project_id

    async def _background_refresh_credentials(
        self,
        credentials: Any,
        credential_cache_key: tuple,
        credential_project_id: str | None,
    ) -> None:
        """
        Refresh credentials in the background without blocking the calling request.

        Called when the token is still valid but nearing expiry (proactive refresh).
        Errors are logged but not raised — the current token is still usable.
        """

View on GitHub (pinned to 77b7c6c40c)

Solutions

1. Provide credentials via vertex_credentials, the GOOGLE_APPLICATION_CREDENTIALS env var, or a service account JSON env var.
2. Verify the credential source is valid and accessible (file exists, JSON parses) in the runtime environment.

When it happens

Trigger: Thrown at litellm/llms/vertex_ai/vertex_llm_base.py:488 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/a9ae118e3c0ed2e8. Report an issue: GitHub.