Configuration

FlaskRMQ.init_app() reads RABBITMQ_CONNECTIONS. The shape mirrors Flask's common extension configuration style: aliases at the top level, connection fields below.

app.config["RABBITMQ_CONNECTIONS"] = {
    "default": {
        "HOST": "rabbitmq.internal",
        "PORT": 5672,
        "VIRTUAL_HOST": "/orders",
        "USER": "orders-service",
        "PASSWORD": os.environ["RABBITMQ_PASSWORD"],
        "HEARTBEAT": 600,
        "BLOCKED_CONNECTION_TIMEOUT": 300,
        "RECONNECT_INITIAL_BACKOFF": 1.0,
        "RECONNECT_MAX_BACKOFF": 30.0,
    },
    "analytics": {
        "HOST": "analytics-rabbitmq.internal",
        "USER": "publisher",
        "PASSWORD": os.environ["ANALYTICS_RABBITMQ_PASSWORD"],
    },
}

Key	Default	Meaning
HOST	required	Broker hostname or IP.
PORT	5672	AMQP TCP port.
VIRTUAL_HOST	/	RabbitMQ virtual host.
USER	guest	Username. Do not use guest remotely in production.
PASSWORD	guest	Password; load it from secrets or environment.
HEARTBEAT	600	Negotiated heartbeat interval in seconds.
BLOCKED_CONNECTION_TIMEOUT	300	Fail a connection blocked by broker resource alarms after this many seconds.
RECONNECT_INITIAL_BACKOFF	1.0	First consumer reconnect delay.
RECONNECT_MAX_BACKOFF	30.0	Exponential-backoff cap.

Lowercase field names are accepted as well, which is convenient when a mapping comes from a configuration service.

Alias resolution

With one configured alias, using= is optional. With two or more, it is mandatory:

audit_producer = Producer(queue="audit", using="analytics")
audit_consumer = Consumer(queue="audit", using="analytics")
get_setup_registry(using="analytics").register(setup_audit)

Omitting it with multiple aliases raises ConfigurationError. This is deliberate: silently selecting a broker is a data-routing bug.

Application context

Registry and connection accessors resolve the current Flask application. Use them inside an app context. Consumer threads created by the CLI automatically receive their own context, so handlers may safely use current_app, Flask-SQLAlchemy, and other context-aware extensions.

When exactly one Flask app has been initialized, producer calls outside a context can resolve it. For multi-app processes, always push an explicit app context.