docker

SignalFx Agent Docker Image

We provide a Docker image at quay.io/signalfx/signalfx-agent. The image is tagged using the same agent version scheme.

If you are using Docker outside of Kubernetes, you can run the agent in a Docker container and still gather metrics on the underlying host by running it with the following flags:

$ docker run \
    --name signalfx-agent \
    --pid host \
    --net host \
    -v /:/hostfs:ro \
    -v /var/run/docker.sock:/var/run/docker.sock:ro \
    -v /etc/signalfx/:/etc/signalfx/:ro \
    quay.io/signalfx/signalfx-agent:<version>

This assumes you have the agent config in the conventional directory (/etc/signalfx) on the root mount namespace. If you want to use a default, built-in configuration, omit the volume bind mount for the /etc/signalfx directory and see Configuration.

If you have the Docker API available through the conventional UNIX domain socket, you should mount that in to be able to use the docker-container-stats monitor.

It is necessary to mount in the host root filesystem at /hostfs in order to get disk usage metrics for the host filesystems using the collectd/df monitor. You will need to set the hostFSPath: /hostfs config option on that monitor to make it use this non-default path.

The other special config you will need is the etcPath: /hostfs/etc option under the host-metadata monitor config. This tells it where to find certain files like /etc/os-release that are used to generate host metadata such as the Linux distro and version.

By using the --pid host flag, the /proc filesystem in the container will match the host's, so that no special configuration of the /proc path is required.

You may also want to use the Docker observer to automatically discover other containers running in the same Docker engine.

Configuration

For any non-trivial use-cases you will need to provide a custom configuration file, but for simple setups or demo purposes, you can use the supplied agent configuration file in the image by omitting the -v /etc/signalfx:/etc/signalfx/:ro flag in the run command above. Then you can set the following environment variables on the agent container:

Environment Variable	Required	Description
SFX_ACCESS_TOKEN	yes	The SignalFx API access token.
SFX_INGEST_URL	no	Often used in conjunction with the SignalFx Gateway to specify a different target URL for datapoints and events. If not specified, this defaults to the global SignalFx ingest server.
SFX_API_URL	no	If you are operating in a different SignalFx realm, this value will need to be set to the SignalFx API server URL in your realm.

The supplied configuration will also load any additional yaml files found in /etc/signalfx/monitors/ as part of the monitors list in the agent config like so:

monitors:
  - {"#from": "/etc/signalfx/monitors/*.yaml", flatten: true, optional: true}
  - type: collectd/cpu
  ...

For example, you can add an ElasticSearch monitor to a configuration by providing a Docker volume mount to /etc/signalfx/monitors/ with a file that follows the monitor config schema:

- collectd/elasticsearch
  host: localhost
  port: 9200

Other options could be specified according to the ElasticSearch Monitor configuration