# Fluentd & Prometheus
## Installation
Install prometheus plugin using gem:

gem install 'fluent-plugin-prometheus'

## Prometheus plugins

1. **prometheus input plugin**: This plugin provides a HTTP endpoint that exposes metrics to Prometheus server on port 24231.
2. **prometheus_monitor input plugin**: This plugin collects internal metrics in Fluentd. This plugin is used to monitor Fluentd itself.
3. **prometheus_output_monitor input plugin**: This plugin collects internal metrics for **output plugin** in Fluentd.

## Configuration

Configure a prometheus metrics so that Prometheus can scrape metrics.

<source>
  @type prometheus
</source>

Next, create a message that will be transformed to prometheus metrics. In the below example, we will create a message that shows the current CPU usage.

<source>
  @type exec
  tag mcnlab
  command grep 'cpu ' /proc/stat | awk '{print usage=($2+$4)*100/($2+$4+$5)}'
  <parse>
    keys usage
  </parse>
  run_interval 1s
</source>

Then, create a Prometheus metrics from the message:

<filter mcnlab>
  @type prometheus
  <metric>
    name mcnlab_cpu_usage
    type gauge
    desc The total number of CPU in message.
    key usage
  </metric>
</filter>

Finally, copy the message to stdout

<match mcnlab>
  @type copy
  <store>
    @type stdout
  </store>
</match>

## Sample Configuration file

<source>
  @type prometheus
</source>
<source>
  @type exec
  tag mcnlab
  command grep 'cpu ' /proc/stat | awk '{print usage=($2+$4)*100/($2+$4+$5)}'
  <parse>
    keys usage
  </parse>
  run_interval 1s
</source>

<filter mcnlab>
  @type typecast
  types usage:float
</filter>

<filter mcnlab>
  @type prometheus
  <metric>
    name mcnlab_cpu_usage
    type gauge
    desc The total number of cpu usage in percentage
    key usage
  </metric>
</filter>

<match mcnlab>
  @type copy
  <store>
    @type stdout
  </store>
</match>