Telegraf allows users to specify multiple output sinks in the configuration file.
Note: Telegraf plugins added in the current release are noted with
-- NEW in v1.6. The Release notes and changelog has a list of new plugins and updates for other plugins. See the plugin README files for more details.
Supported Telegraf output plugins
If the point value being sent cannot be converted to a float64, the metric is skipped.
Metrics are grouped by converting any
_ characters to
. in the Point Name.
Metrics are written to a topic exchange using
tag, defined in configuration file as
RoutingTag, as a routing key.
The CloudWatch (
cloudwatch) output plugin send metrics to Amazon CloudWatch.
The Discard (
discard) output plugin simply drops all metrics that are sent to it. It is only meant to be used for testing purposes.
The File (
file) output plugin writes Telegraf metrics to files.
The Graylog (
graylog) output plugin writes to a Graylog instance using the
The InfluxDB (
influxdb) output plugin writes to InfluxDB using HTTP or UDP.
Instrumental accepts stats in a format very close to Graphite, with the only difference being that the type of stat (gauge, increment) is the first token, separated from the metric itself by whitespace. The increment type is only used if the metric comes in as a counter through [[input.statsd]].
The Kinesis (
kinesis) output plugin is an experimental plugin that is still in the early stages of development. It will batch up all of the Points in one Put request to Kinesis. This should save the number of API requests by a considerable level.
The NATS Output (
nats) output plugin writes to a (list of) specified NATS instance(s).
The NSQ (
nsq) output plugin writes to a specified NSQD instance, usually local to the producer. It requires a server name and a topic name.
The OpenTSDB (
opentsdb) output plugin writes to an OpenTSDB instance using either the telnet or HTTP mode.
Using the HTTP API is the recommended way of writing metrics since OpenTSDB 2.0 To use HTTP mode, set
useHttp to true in config. You can also control how many metrics are sent in each HTTP request by setting
batchSize in config. See http://opentsdb.net/docs/build/html/api_http/put.html for details.
The Prometheus Client (
prometheus_client) output plugin starts a Prometheus Client, it exposes all metrics on
/metrics (default) to be polled by a Prometheus server.
The Wavefront (
wavefront) output plugin writes to a Wavefront proxy, in Wavefront data format over TCP.
Deprecated Telegraf output plugins
The Riemann Legacy (
riemann_legacy) output plugin will be deprecated in a future release, see https://github.com/influxdata/telegraf/issues/1878 for more details & discussion.