HumanSignal/label-studio · error · KeyError

Container not found: {self.container}

Error message

Container not found: {self.container}

What it means

During Azure Blob connection validation, the SDK raises ResourceNotFoundError when the target container cannot be found, which the code translates to KeyError('Container not found: {name}'). This means credentials/auth succeeded but the named container does not exist (or is inaccessible) in the account.

Source

Thrown at label_studio/io_storages/azure_blob/models.py:90

            + ';EndpointSuffix=core.windows.net'
        )
        client = BlobServiceClient.from_connection_string(conn_str=connection_string)
        container = client.get_container_client(str(self.container))
        return client, container

    def get_container(self):
        _, container = self.get_client_and_container()
        return container

    def validate_connection(self, **kwargs):
        logger.debug('Validating Azure Blob Storage connection')
        client, container = self.get_client_and_container()

        try:
            container_properties = container.get_container_properties()
            logger.debug(f'Container exists: {container_properties.name}')
        except ResourceNotFoundError:
            raise KeyError(f'Container not found: {self.container}')

        # Check path existence for Import storages only
        if self.prefix and 'Export' not in self.__class__.__name__:
            logger.debug(f'Test connection to container {self.container} with prefix {self.prefix}')
            prefix = str(self.prefix)
            try:
                blob = next(container.list_blob_names(name_starts_with=prefix))
            except StopIteration:
                blob = None

            if not blob:
                raise KeyError(f'{self.url_scheme}://{self.container}/{self.prefix} not found.')

    def get_bytes_stream(self, uri, range_header=None):
        """Get file bytes from Azure Blob storage as a streaming object with metadata.

        Implements range request support similar to GCS and S3 implementations:
        - Accepts ``range_header`` in format ``bytes=start-end``

View on GitHub (pinned to 0b49e9b539)

Solutions

1. Run 'az storage container list --account-name <acct>' and use an exact existing container name
2. Create the container if it should exist: az storage container create --name <name>
3. Verify account_name/key refer to the correct storage account (case-sensitive container names)
4. Check access key / SAS permissions allow reading container metadata

Example fix

// before
{"container": "Documents"}  // actual container is 'documents' → KeyError Container not found
// after
{"container": "documents"}  → 200 valid

from azure.storage.blob import BlobServiceClient
svc = BlobServiceClient(account_url=f"https://{acct}.blob.core.windows.net", credential=key)
existing = {c.name for c in svc.list_containers()}
assert container_name in existing, f"Container '{container_name}' does not exist; have: {existing}"

def container_exists(svc, name):
    return any(c.name == name for c in svc.list_containers())

try:
    resp = requests.post(f"{LS_URL}/api/storages/azure/validate/", json=payload, headers=headers)
    resp.raise_for_status()
except requests.HTTPError as e:
    if "Container not found" in resp.text:
        logging.error("Create the container or fix its name (case-sensitive): %s", resp.text)

When it happens

Trigger: POST /api/storages/azure/validate or creating a storage where 'container' names a container that does not exist in the account, or the account key lacks rights to read container properties.

Common situations: Typo in container name or case mismatch; container created in a different storage account than the configured one; container deleted after being configured; restricted IAM denying get_container_properties.

Related errors

• {self.url_scheme}://{self.container}/{self.prefix} not found
• {underlying connection validation error}
• {connection validation error}
• Validation failed on {}: {}
• Label config contains non-unique names:

AI-assisted analysis of HumanSignal/label-studio@0b49e9b539 (2026-08-29). Data as JSON: /api/errors/658e8bfeb7cee9af. Report an issue: GitHub.