contactkey

Contactkey is a Go command-line tool that orchestrate deployments on complex environments.

Deployment Flow

Deployers, Sources, Binaries, Hooks, Lock System

Deployers

• GGN
• Kubernetes (to do)

Sources (upstream application code)

• Stash
• Github (to do)

Binaries (container images)

• Nexus

Hooks

• Slack
• ExecCommand
• NewRelic
• MailSender (to do)

Lock System

• FileLock
• Etcd (to do)
• Redis (to do)

Commands

cck deploy environment service    Deploy the service in an environment
cck diff environment service      Diff between what's currently deployed and what's going to be deployed (Sources)
cck list environment service      List versions of the service in an environment
cck rollback environment service  Rollback the service in an environment

Global configuration file for the cck command.

The configuration file is located at ~/.contackey/config.yml

workPath: /tmp/manifests   # Location of services manifest

screenMandatory: true      # Check if the user is launching cck in a screen/tmux (not mandatory)

globalEnvironments:        # Define the cck environment. It can be anything.
  - preprod                # It will be used as the cck environment for
  - prod                   # the command line.

deployers:                 # Define the various deployers used in service manifest.
  ggn:                     # This will enable the ggn deployment feature
    vcsRegexp: -v(.+)      # Extract the vcs sha1 from pod version. (Not mandatory)
    workPath: /tmp
    environments:          # Link between ggn environment and cck environment created above.
      preprod: staging
      prod:    production
   k8s:                    # This will enable deployment on k8s using helm
    vcsRegexp: -v(.+)      # Extract the vcs sha1 from pod version. (Not mandatory)
    workPath: /tmp
    environments:          # Link between cck environments and k8s clusters
      local:
        cluster: ops
      preprod:
        cluster: preprod

sources:      # Define various version control system used in service manifest.
  stash:                   # Currently we have only Stash supported.
    user:        user
    password:    password
    url:         url
    sha1MaxSize: 2         # Cut a sha1 if it's too long. (E.g: abcd => ab) (Not mendatory)

binaries:         # Define various repository manager used in service manifest.
  nexus:                   # Currently we have only Nexus supported.
    url:        127.0.0.1
    repository: repository
    group:      group

hooks:                     # Define various hooks used in service manifest.
  slack:                   # Currently we have only Slack supported.
    url:   127.0.0.1
    token: token

  newRelic:
    url: https://api.newrelic.com
    apiKey: "<your_api_key>"

metrics:
  prometheus:              # Push internal metrics to a prometheus gateway
    url: http://demo.robustperception.io:9091

lockSystem:                # Define a lock system in order to avoid multiple command launch. (Not mandatory)
  fileLock:                # Currently we only have the lock by file.
    filePath: /tmp         # Path where the is going to be written.

potentialUsername:       # By default get the current user.
  - USER                 # You can also have a list of env variable where you can find the username.
  - bamboo.jira.username # It stops on the first username found in the list.

Configuration by service (manifest)

One file per service. They must be located in the workPath defined above. The file name will be service name in the cck command.

deployment:                # Define the deployment type we are going to use
  ggn:                     # for this service.
    pod: pod-webhooks      # podName used in ggn
    service: webhooks      # service used in ggn
sources:
  stash:                   # Version control system used for the service. (Only one)
    branch: master         # The default "stable" branch we usually deploy
    repository: repository
    project: project
binaries:
  nexus:                   # Repository manager used for the service. (Only one)
    artifact: pod-webhooks
hooks:                     # Hooks we are going to call before and after.
  slack:                   # the deployment. (You can have several one)
    channel:   channel
    stopOnError: false     # If an error occurs stop the deployment process (not mandatory, default: false)
  newRelic:
    applicationId: "<id>"
    stopOnError: false     # If an error occurs stop the deployment process (not mandatory, default: false)
  execCommand:             # Execute a command before and after the deployment process.
    onInit:                # Execute a command after reading the manifest
      - { command: "ls", args: ["-lah"] } # args is not mandatory
    onPredeploy:           # Execute a command before the deployment
      - { command: "ls", args: ["-lah"] } # args is not mandatory
    onPostDeploy:          # Execute a command after the deployment
      - { command: "cd /tmp"}
    stopOnError: true      # If an error occurs stop the deployment process (not mandatory default false)

There are multiple deployers, the above example uses ggn, but you can deploy to kubernetes as well. Here is an example : k8s-manifest.sample.yaml.