Skip to content

Exporting Metrics to Graphite via the collectd Write Plugin

This document provides an example of using the write_graphite write plugin to export metrics to Graphite.

Example Workflow

Example of metric

This example shows how to work with the single curl_json-wallarm_nginx/gauge-abnormal metric, which shows the number of requests processed by the filter node.

Example workflow

The following deployment scheme is used in this document:

  • Wallarm filter node is deployed on a host accessible via the 10.0.30.5 IP address and the node.example.local fully qualified domain name.

    The write_graphite plugin for collectd on the filter node is configured as follows:

    • All metrics are sent to the 10.0.30.30 server listening on the 2003/TCP port.
    • Some Wallarm-specific collectd plugins support multiple instances, so the write_graphite plugin contains the SeparateInstances parameter set to true. The true value means that the plugin can work with several instances.

    A complete list of plugin options is available here.

  • Both graphite and grafana services are deployed as Docker containers on a separate host with the 10.0.30.30 IP address.

    The graphite service with Graphite is configured as follows:

    • It listens for incoming connections on the 2003/TCP port, to which collectd will send the filter node metrics.
    • It listens for incoming connections on the 8080/TCP port, through which communication with Grafana will occur.
    • The service shares the sample-net Docker network with the grafana service.

    The grafana service with Grafana is configured as follows:

    • The Grafana web console is available at http://10.0.30.30:3000.
    • The service shares the sample-net Docker network with the graphite service.

Configuring Metrics Export to Graphite

Prerequisites

It is assumed that

  • Docker Community Edition and docker-compose are already installed on the 10.0.30.30 Docker host.
  • The node.example.local filter node is already deployed, configured, available for further configuration (for example, via the SSH protocol), and working.

Deploying Graphite and Grafana

Deploy Graphite and Grafana on the Docker host:

  1. Create a docker-compose.yaml file with the following content:

    version: "3"

services:
  grafana:
    image: grafana/grafana
    container_name: grafana
    restart: always
    ports:
      - 3000:3000
    networks:
      - sample-net

  graphite:
    image: graphiteapp/graphite-statsd
    container_name: graphite
    restart: always
    ports:
      - 8080:8080
      - 2003:2003
    networks:
      - sample-net

networks:
  sample-net:

  2. Build the services by executing the docker-compose build command.

  3. Run the services by executing the docker-compose up -d graphite grafana command.

At this point, you should have Graphite running and ready to receive metrics from collectd, and Grafana ready to monitor and visualize the data stored in Graphite.

Configuring collectd

Configure collectd to download metrics to Graphite:

  1. Connect to the filter node (for example, using the SSH protocol). Make sure you are logged in as root or another account with superuser privileges.

  2. Create a file named /etc/collectd/collectd.conf.d/export-to-graphite.conf with the following content:

    LoadPlugin write_graphite

<Plugin write_graphite>
 <Node "node.example.local">
   Host "10.0.30.30"
   Port "2003"
   Protocol "tcp"
   SeparateInstances true
 </Node>
</Plugin>

    The following entities are configured here:

    1. The host name from which metrics are collected (node.example.local).
    2. The server to which metrics should be sent (10.0.30.30).
    3. The server port (2003) and the protocol (tcp).
    4. The data transfer logic: the data of one instance of the plugin is separated from the data of another instance (SeparateInstances true).

  3. Restart the collectd service by running the appropriate command:

    sudo service collectd restart

    sudo systemctl restart collectd

Now Graphite will receive all metrics of the filter node. You can visualize the metrics you are interested in, and monitor them with Grafana.