microsoft/semantic-kernel · error · MemoryConnectorInitializationError

Failed to create MongoDB Atlas settings.

Error message

Failed to create MongoDB Atlas settings.

What it means

Raised as a MemoryConnectorInitializationError in MongoDBAtlasMemoryStore.__init__ when MongoDBAtlasSettings construction raises a ValidationError. The settings object validates required fields (connection string, database name, index name) from constructor args and/or environment variables; any missing or invalid required field triggers the pydantic ValidationError which is then wrapped here.

Source

Thrown at python/semantic_kernel/connectors/memory_stores/mongodb_atlas/mongodb_atlas_memory_store.py:70

            connection_string (str): The connection string for the MongoDB Atlas instance.
            database_name (str): The name of the database.
            read_preference (ReadPreference): The read preference for the connection.
            env_file_path (str): The path to the .env file containing the connection string.
            env_file_encoding (str): The encoding of the .env file.

        """
        from semantic_kernel.connectors.mongodb import MongoDBAtlasSettings

        try:
            mongodb_settings = MongoDBAtlasSettings(
                database_name=database_name,
                index_name=index_name,
                connection_string=connection_string,
                env_file_path=env_file_path,
                env_file_encoding=env_file_encoding,
            )
        except ValidationError as ex:
            raise MemoryConnectorInitializationError("Failed to create MongoDB Atlas settings.") from ex

        self.mongo_client: motor_asyncio.AsyncIOMotorClient = motor_asyncio.AsyncIOMotorClient(
            mongodb_settings.connection_string.get_secret_value(),
            read_preference=read_preference,
            driver=DriverInfo("Microsoft Semantic Kernel", metadata.version("semantic-kernel")),
        )
        self.database_name: str = mongodb_settings.database_name
        self.index_name: str = mongodb_settings.index_name

    @property
    def database(self) -> core.AgnosticDatabase:
        """The database object."""
        return self.mongo_client[self.database_name]

    @property
    def num_candidates(self) -> int:
        """The number of candidates to return."""
        return self.__num_candidates

View on GitHub (pinned to c028a0c7dc)

Solutions

1. Set the required environment variables: MONGODB_ATLAS_CONNECTION_STRING, MONGODB_ATLAS_DATABASE_NAME, and the index name, or pass them as constructor args.
2. If using a .env file, verify env_file_path is correct and the file is readable.
3. Inspect the chained ValidationError (__cause__) for the specific missing/invalid field.
4. Use MongoDBAtlasSettings() standalone first to surface detailed field-level errors before constructing the store.

Example fix

// before
store = MongoDBAtlasMemoryStore()  # MemoryConnectorInitializationError (missing settings)
// after
# .env
# MONGODB_ATLAS_CONNECTION_STRING=mongodb+srv://...
# MONGODB_ATLAS_DATABASE_NAME=mydb
store = MongoDBAtlasMemoryStore(index_name='vector_index')

from semantic_kernel.connectors.mongodb import MongoDBAtlasSettings

def validate_mongodb_settings(**kwargs):
    try:
        return MongoDBAtlasSettings(**kwargs)
    except Exception as e:
        raise ValueError(f'MongoDB Atlas settings invalid: {e}') from e

settings = validate_mongodb_settings(
    database_name='mydb', index_name='vector_index', connection_string=conn_str
)

from semantic_kernel.exceptions.memory_connector_exceptions import MemoryConnectorInitializationError

try:
    store = MongoDBAtlasMemoryStore(index_name='vector_index')
except MemoryConnectorInitializationError as e:
    logging.error('MongoDB Atlas settings invalid: %s', e.__cause__)
    raise

When it happens

Trigger: Constructing MongoDBAtlasMemoryStore without MONGODB_ATLAS_CONNECTION_STRING / database name / index name in the environment or args. Passing an invalid connection string format. A .env file that is missing or unreadable at the given env_file_path.

Common situations: Missing .env file or MONGODB_ATLAS_* env vars in the deployment environment. Typo in the env var name. Connection string lacking required MongoDB Atlas vector search parameters. Running locally without loading the env file.

Related errors

• Failed to validate Vertex AI settings: {e}
• Failed to validate Vertex AI settings: {e}
• Failed to validate Vertex AI settings: {e}
• Failed to create MistralAI settings.
• The OpenAI text model ID is required.

AI-assisted analysis of microsoft/semantic-kernel@c028a0c7dc (2026-08-13). Data as JSON: /api/errors/35ad25e8d61a7fd7. Report an issue: GitHub.