That's what we'll solve, let's load the model before the requests are handled, but only right before the application starts receiving requests, not while  the code is being loaded.

## Lifespan { #lifespan }

authors = [
    { name = "Sebastián Ramírez", email = "******@****.***" },
]
classifiers = [
    "Intended Audience :: Information Technology",
    "Intended Audience :: System Administrators",
    "Operating System :: OS Independent",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python",
    "Topic :: Internet",
    "Topic :: Software Development :: Libraries :: Application Frameworks",

    mod.engine.dispose()


def test_crud_app(client: TestClient):
    # TODO: this warns that SQLModel.from_orm is deprecated in Pydantic v1, refactor
    # this if using obj.model_validate becomes independent of Pydantic v2
    with warnings.catch_warnings(record=True):
        warnings.simplefilter("always")
        # No heroes before creating
        response = client.get("heroes/")

- [Configuring etcd](https://github.com/minio/minio/blob/master/docs/sts/etcd.md)

### Setup MinIO with Identity Provider

Make sure we have followed the previous step and configured each software independently, once done we can now proceed to use MinIO STS API and MinIO server to use these credentials to perform object API operations.

#### KeyCloak

```
export MINIO_ROOT_USER=minio
export MINIO_ROOT_PASSWORD=minio123

Now let's go back a bit and understand what is all that.

The `password` "flow" is one of the ways ("flows") defined in OAuth2, to handle security and authentication.

OAuth2 was designed so that the backend or API could be independent of the server that authenticates the user.

But in this case, the same **FastAPI** application will handle the API and the authentication.

So, let's review it from that simplified point of view:

        assertNotNull(results);
    }

    @Test
    public void testMultipleConverterInstances() throws IOException {
        // Test that multiple converter instances work independently
        AnalyzerConverter converter1 = new AnalyzerConverter(client, settings);
        AnalyzerConverter converter2 = new AnalyzerConverter(client, settings);

        String text = "test";

* Hashage de **mots de passe sécurisé** par défaut.
* Authentification par **jetons JWT**.
* Modèles **SQLAlchemy** (indépendants des extensions Flask, afin qu'ils puissent être utilisés directement avec des *workers* Celery).
* Modèle de démarrages basiques pour les utilisateurs (à modifier et supprimer au besoin).
* Migrations **Alembic**.

Eso es lo que resolveremos, vamos a cargar el modelo antes de que los requests sean manejados, pero solo justo antes de que la aplicación comience a recibir requests, no mientras el código se está cargando.

        assertTrue(result.length() > "TransWaitNamedPipeResponse[]".length());
    }

    @Test
    @DisplayName("Multiple instances should be independent")
    void testMultipleInstances() {
        // Arrange
        TransWaitNamedPipeResponse response1 = new TransWaitNamedPipeResponse(mockConfig);

         * The kerberos session key is not accessible via the JGSS API. IBM and
         * Oracle both implement a similar API to make an ExtendedGSSContext
         * available. That API is accessed via reflection to make this independent
         * of the runtime JRE
         */
        if (EXT_GSS_CONTEXT_CLASS == null || INQUIRE_SEC_CONTEXT == null || INQUIRE_TYPE_SESSION_KEY == null) {