Skip to content
This repository has been archived by the owner on Apr 21, 2020. It is now read-only.

newrelic/check_docker

Archived header

check_docker

A Nagios check to check some basic statistics reported by the Docker daemon. Additionally validates the absence of Ghost containers and may optionally be made to require the presence of a container running from a particular image tag.

check_docker is written in Go and is multi-threaded to keep the drag time low on your Nagios server. It makes two API requests to the Docker daemon, one to /info and one to /containers/json and processes the results, all simultaneously.

It is built using the the go_nagios framework.

Installing

If you would rather not build the binaries yourself, you can install compiled, statically-linked binaries for Linux or MacOSX. Simply download the tarball, extract it and use the binary of your choice.

Building

go get github.com/newrelic/go_nagios
go build

Usage

Usage of ./check_docker:
  -base-url="http://docker-server:2375": The Base URL for the Docker server
  -container-name="": The name of a container that must be running on the Docker server
  -warn-data-space=100: Warning threshold for Data Space
  -crit-data-space=100: Critical threshold for Data Space
  -warn-meta-space=100: Warning threshold for Metadata Space
  -crit-meta-space=100: Critical threshold for Metadata Space
  -image-id="": An image ID that must be running on the Docker server
  -tls-cert="": Path to TLS cert file
  -tls-key="": Path to TLS key file
  -tls-ca="": Path to TLS CA file

-base-url: Here you specify the base url of the docker server.

-container-name: Allows you to specify the name of a container that should be running on the server.

-image-id: You can specify an image tag that needs to be running on the server for certain cases where you have pegged a container to a server (e.g. each server has a Nagios monitoring container running to report on server health). Will not require any particular image if left off.

-(warn|crit)-(meta|data)-space: the thresholds at which the named Nagios status codes should be emitted. These are percentages, so -crit-data-space=95 would send a CRITICAL response when the threshold of 95% is crossed. Defaults are 100%.

Contributions

Contributions are more than welcome. Bug reports with specific reproduction steps are great. If you have a code contribution you'd like to make, open a pull request with suggested code.

Pull requests should:

  • Clearly state their intent in the title
  • Have a description that explains the need for the changes
  • Include tests!
  • Not break the public API

Testing for Contributors

go test can be run on 2 different platforms:

  1. In Darwin(aufs), assuming you already setup Boot2Docker:

    docker run -t -d --name named_container busybox:latest
    export DOCKER_IMAGE=busybox:latest
    export DOCKER_CONTAINER_NAME=named_container
    
    cd $GOPATH/src/github.com/newrelic/check_docker
    go get ./... && go test
    
  2. In Linux(devicemapper), you are running the tests inside vagrant:

    cd $GOPATH/src/github.com/newrelic/check_docker
    vagrant up --provider virtualbox
    vagrant ssh
    
    # Inside Vagrant
    sudo docker run -t -d --name named_container busybox:latest
    export DOCKER_IMAGE=busybox:latest
    export DOCKER_CONTAINER_NAME=named_container
    
    export GOPATH=/go
    cd $GOPATH/src/github.com/newrelic/check_docker
    go get ./... && go test
    

Copyright (c) 2014 New Relic, Inc. All rights reserved.