influxdb_exporter

README

InfluxDB Exporter

An exporter for metrics in the InfluxDB format used since 0.9.0. It collects metrics in the line protocol via a HTTP API, transforms them and exposes them for consumption by Prometheus.

This exporter supports float, int and boolean fields. Tags are converted to Prometheus labels.

The exporter also listens on a UDP socket, port 9122 by default.

Timestamps

By default metrics exposed without original timestamps like this:

http_requests_total{method="post",code="200"} 1027
http_requests_total{method="post",code="400"}    3

If you want to add original timestamps to exposed metrics, please use flag --timestamps and metrics will looks like:

http_requests_total{method="post",code="200"} 1027 1395066363000
http_requests_total{method="post",code="400"}    3 1395066363000

When querying, this means that the sample is attributed to the time it was submitted to the exporter, not the time Prometheus scraped it. However, if the metric was submitted multiple time in between exporter scrapes, only the last value and timestamp will be stored.

Alternatives

If you are sending data to InfluxDB in Graphite or Collectd formats, see the graphite_exporter and collectd_exporter respectively.

This exporter is useful for exporting metrics from existing collectd setups, as well as for metrics which are not covered by the core Prometheus exporters such as the Node Exporter.

The exporter acts like an InfluxDB server, it does not connect to one. For metrics concerning the InfluxDB server, use the metrics endpoint built into InfluxDB.

If you are already using Telegraf, it can serve the same purpose as this exporter with the outputs.prometheus_client plugin.

For more information on integrating between the Prometheus and InfluxDB ecosystems, see the influxdata integration page.

Example usage with Telegraf

The influxdb_exporter appears as a normal InfluxDB server. To use with Telegraf for example, put the following in your telegraf.conf:

[[outputs.influxdb]]
  urls = ["http://localhost:9122"]

Or if you want to use UDP instead:

[[outputs.influxdb]]
  urls = ["udp://localhost:9122"]

Note that Telegraf already supports outputing Prometheus metrics over HTTP via outputs.prometheus_client, which avoids having to also run the influxdb_exporter.