The end of a pipeline passes on the data to the specified output. Usually the whole event is passed by default, but there are often options like input-field and body to specify what data to send.

Many of these outputs support retrying:

        count: 3 # number of times to retry
        pause: 2s # pause between retries
        forever: false # instead of count, just keep rerying

While data is waiting to be sent in these outputs, it will be buffered until it can be sent.

NOTE the buffer is currently unbounded so watch out for possible memory exhaustion.

Most outputs can also be called as actions with action output, with some restrictions. The servers http-server, tcp (with server: true) make no sense as actions, and batching is not allowed. The output is sent and the incoming event passed on unaltered. This can be used to send events to multiple outputs.

write (output)Write to a file or `console` (a special name for standard output)
tcp (output)Send data to a TCP server
udp (output)Send data to a UDP server
exec (output)Execute arbitrary commands
elastic (output)Send events to Elasticsearch 5 server
amqp (output)Send events to AMQP server
redis (output)Write to Redis in-memory key-value store
http-server (output)Run HTTP server
http-post (output)Run POST HTTP queries