BerriAI/litellm · error · HTTPException

Error caching temporary mcp server: {e}

Error message

Error caching temporary mcp server: {e}

What it means

Returned (500) when the temporary-server setup block fails: either _persist_draft_mcp_server (DB write of the draft row) or _cache_temporary_mcp_server_in_redis (Redis write with TEMPORARY_MCP_SERVER_TTL_SECONDS). The original exception is logged via verbose_proxy_logger.exception, then surfaced as this generic 500; nothing usable is returned to the caller.

Source

Thrown at litellm/proxy/management_endpoints/mcp_management_endpoints.py:1670

                temp_record,
                credentials_are_encrypted=False,
            )
            _cache_temporary_mcp_server(
                temporary_server,
                ttl_seconds=TEMPORARY_MCP_SERVER_TTL_SECONDS,
            )
            await _persist_draft_mcp_server(
                payload_with_credentials,
                temp_record.server_id,
                created_by,
            )
            await _cache_temporary_mcp_server_in_redis(
                temporary_server,
                ttl_seconds=TEMPORARY_MCP_SERVER_TTL_SECONDS,
            )
        except Exception as e:
            verbose_proxy_logger.exception("Error caching temporary mcp server: %s", e)
            raise HTTPException(
                status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
                detail={"error": f"Error caching temporary mcp server: {e}"},
            )

        return _redact_mcp_credentials(temp_record)

    async def _mcp_oauth_user_api_key_auth(request: Request) -> UserAPIKeyAuth:
        """
        Auth dependency for MCP OAuth browser-navigation endpoints (/authorize, /token).

        Tries the Authorization header first. Falls back to decoding the UI
        'token' session cookie (set by SSO login) to extract the API key, which
        allows browser-based OAuth redirects to work without an explicit
        Authorization header.
        """
        import jwt as _jwt

        from litellm.proxy.proxy_server import master_key

View on GitHub (pinned to 77b7c6c40c)

Solutions

1. Check the proxy logs for the wrapped exception's traceback to see which leg failed (draft persist vs Redis cache).
2. Verify the proxy's Redis connection (redis_url) and that Redis is reachable.
3. Verify DB connectivity/migrations if the draft persist is the failing call.
4. Retry the session flow once infrastructure is healthy; temp entries are TTL-scoped so partial state ages out.

try:
    import redis
    redis.Redis.from_url(REDIS_URL).ping()
except Exception as e:
    raise RuntimeError(f"Redis unavailable; temp MCP server flow will 500: {e}")

for attempt in range(2):
    try:
        return create_temp_session(payload)
    except HTTPError as e:
        if e.response.status_code == 500 and attempt == 0:
            sleep(2)
            continue  # transient Redis/DB hiccup; temp entries are TTL-scoped so retry is safe
        raise

When it happens

Trigger: Running the admin MCP OAuth session flow while Redis is down or redis_url is not configured on the proxy; the draft DB write failing (DB down, schema drift); values in the payload failing serialization on the way into the cache.

Common situations: Proxy deployed without the Redis instance that temporary MCP server caching requires; Redis restarted or evicting mid-flow; DB outage during OAuth setup.

Related errors

• Error fetching cache settings: {e}
• Error updating cache settings: {e}
• Error creating mcp server: {e}
• User does not have permission to create temporary mcp server
• MCP server {server_id} not found

AI-assisted analysis of BerriAI/litellm@77b7c6c40c (2026-08-18). Data as JSON: /api/errors/c3f42aa244fe7ede. Report an issue: GitHub.