microsoft/semantic-kernel · error · VectorStoreInitializationException

The connection string is missing.

Error message

The connection string is missing.

What it means

After successfully constructing MongoDBAtlasSettings (validation passed), the constructor checks that a connection string is actually present. If the settings object exists but `connection_string` is falsy (empty or None), this VectorStoreInitializationException is raised. It indicates that settings validation did not fail, but the connection string specifically was not provided — a distinct, more targeted message than the general settings failure.

Source

Thrown at python/semantic_kernel/connectors/mongodb.py:585

                managed_client=managed_client,
                database_name=database_name or DEFAULT_DB_NAME,
                embedding_generator=embedding_generator,
            )
            if callable(mongo_client.append_metadata):
                mongo_client.append_metadata(DRIVER_METADATA)
            return

        try:
            mongodb_atlas_settings = MongoDBAtlasSettings(
                env_file_path=env_file_path,
                connection_string=connection_string,
                database_name=database_name,
                env_file_encoding=env_file_encoding,
            )
        except ValidationError as exc:
            raise VectorStoreInitializationException("Failed to create MongoDB Atlas settings.") from exc
        if not mongodb_atlas_settings.connection_string:
            raise VectorStoreInitializationException("The connection string is missing.")

        mongo_client = AsyncMongoClient(
            mongodb_atlas_settings.connection_string.get_secret_value(),
            driver=DRIVER_METADATA,
        )

        super().__init__(
            mongo_client=mongo_client,
            managed_client=managed_client,
            database_name=mongodb_atlas_settings.database_name,
            embedding_generator=embedding_generator,
        )

    @override
    def get_collection(
        self,
        record_type: type[TModel],
        *,

View on GitHub (pinned to c028a0c7dc)

Solutions

1. Explicitly provide `connection_string` to the MongoDBAtlasStore constructor.
2. Ensure the `MONGODB_ATLAS_CONNECTION_STRING` environment variable is set to a non-empty value.
3. Check the `.env` file for a blank or whitespace-only connection string value.
4. Validate the connection string is non-empty before constructing the store.

Example fix

// before (env var set to empty string)
store = MongoDBAtlasStore()
// after
store = MongoDBAtlasStore(connection_string="mongodb+srv://user:pass@cluster.mongodb.net")

import os
conn = os.getenv("MONGODB_ATLAS_CONNECTION_STRING", "")
if not conn or not conn.strip():
    raise ValueError("Connection string is empty or missing; set MONGODB_ATLAS_CONNECTION_STRING.")

try:
    store = MongoDBAtlasStore()
except VectorStoreInitializationException as e:
    if "connection string is missing" in str(e):
        # explicitly pass connection_string argument
        ...

When it happens

Trigger: MongoDBAtlasSettings builds without a ValidationError (e.g. connection_string is an optional field), but the connection_string ends up empty/None. This can happen if connection_string is conditionally required and the settings model allows it to be absent.

Common situations: The connection string env var is set to an empty string. A custom settings subclass or env configuration makes connection_string optional at the Pydantic level. The `.env` file exists but the connection string line is blank or commented.

Related errors

• Failed to create Azure OpenAI settings: {exc}
• Please provide either an api_key, ad_token, ad_token_provide
• Please provide an Azure OpenAI endpoint
• Please provide an Azure OpenAI Responses deployment name
• Configuration not found, please setup the notebooks first us

AI-assisted analysis of microsoft/semantic-kernel@c028a0c7dc (2026-08-13). Data as JSON: /api/errors/210d23de6c181e82. Report an issue: GitHub.