Allow to have a river plugin that automatically index changes by listening to a rabbitmq queue. The format of the message is the same as the bulk API (#371), here is an example:

{ "index" : { "index" : "test", "type" : "type1", "id" : "1" }
{ "type1" : { "field1" : "value1" } }
{ "delete" : { "index" : "test", "type" : "type1", "id" : "2" } }
{ "create" : { "index" : "test", "type" : "type1", "id" : "1" }
{ "type1" : { "field1" : "value1" } }

Creating a rabbitmq river is simple:

curl -XPUT 'localhost:9200/_river/my_river/_meta' -d '{
    "type" : "rabbitmq",
}'

There are several configuration that can be set along with the type on an rabbitmq element:

• host: The host to connect to, defaults to localhost.
• port: The port to connect to, defaults to the default AMQP port.
• user: The user name. Default to not being set.
• password: The password. Defaults to not being set.
• vhost: The virtual host.
• queue: The queue name. Defaults to elasticsearch.
• exchange: The exchange name. Defaults to elasticsearch.
• routing_key: The routing key. Defaults to elasticsearch.

On the index element, the following can be controlled:

• bulk_size: The total bulk size when aggregating messages.
• bulk_timeout: One a first message is received, how much time to wait to try and aggregate more messages to a single bulk.