MQTT is a lightweight queueing protocol commonly used in the IoT industry. The Arroyo MQTT connector can use MQTT brokers as sources and sinks using configurable reliability.

​
Configuring the Connection

MQTT connections can be created with the Web UI or SQL DDL.

Creating an MQTT connection

An MQTT connection has several required and optional fields:

FieldDescriptionRequiredExample
urlThe URL of the MQTT brokerYestcp://localhost:1883
usernameUsername for the MQTT brokerFor password authbob
passwordPassword for the MQTT brokerfor password authhunter2
tls.caPath to the CA fileFor TLS auth/mount/certs/ca.pem
tls.certPath to the client cert file authFor TLS auth/mount/certs/client.pem
tls.keyPath to the client key fileFor TLS auth/mount/certs/client.key
topicThe MQTT topic to read or write fromYesevents
qosOne of ‘AtMostOnce’, ‘AtleastOnce’, or ‘ExactlyOnce’NoAtMostOnce
typeEither ‘source’ or ‘sink’Yessource
sink.retainWhether to retain messages published to this topicNofalse

​
Connection Profiles

Arroyo connectors can split their configuration into two parts: the profile config, which covers the common details of how to connect and authenticate against a cluster, and the table config which is specific to a particular instance of the connector.

For MQTT, the profile represents a particular MQTT broker or cluster, while the table represents a particular topic.

MQTT profiles can be defined via the Web UI and shared across multiple tables. They can be used in SQL DDL by setting the connection_profile option in the WITH clause.

See the connection profile docs for more details.

​
Examples

​
MQTT Source

create table mqtt_source (
    id TEXT,
    lat FLOAT,
    lng FLOAT,
    type INT
) with (
    connector = 'mqtt',
    url = 'tcp://localhost:1833',
    type = 'source',
    format = 'json',
    topic = 'events'
);

​
MQTT Sink

create table mqtt_sink with (
    connector = 'mqtt',
    url = 'tcp://localhost:1833',
    type = 'sink',
    format = 'json',
    topic = 'results',
    'sink.retain' = 'true'
);
KinesisMySQL