release is a job that is used to create a semantically versioned object that represents an immutable list of service definitions, in our case a manifest. The benefit of using this job would be if you have multiple services that make up your application and you want them all to be deployed as a unit.

This job can also be used to simply maintain an incrementing version number, without being associated with any particular set of manifests.

Shippable DevOps Assembly Lines store the service definitions in the versions of the following jobs:

  • manifest is a single service definition.
  • deploy is a collection of one or more manifests that have been deployed to a cluster.
  • release is a collection of one or more manifests that have been associated with a version number.

All of the above jobs can be used as INs to create a new release. A new version is created anytime this job is executed.

You can create a release job by adding it to and these jobs execute on Shippable provided shared runtime.

YML Definition

  - name:               <string>                    #required
    type:               release                     #required
    on_start:                                       # optional
      - NOTIFY:         <notification resource name>
      - IN:             <version>                   # optional
      - IN:             <manifest/release/deploy>   # optional
      - IN:             <manifest/release>          # optional
      - IN:             <any job or resource>       # optional
      - TASK:           managed
        bump:           <bump strategies>           # optional, defaults to minor
    on_success:                                 # optional
      - NOTIFY:         <notification resource name>
    on_failure:                                 # optional
      - NOTIFY:         <notification resource name>
    on_cancel:                                      # optional
      - NOTIFY:         <notification resource name>
    always:                                     # optional
      - NOTIFY:     <notification resource name>
  • name -- should be an easy to remember text string

  • type -- is set to manifest

  • steps -- is an object which contains specific instructions to run this Job

    • IN -- You need at least one source of a semVer string, and one manifest-based input (manifest, release, or deploy). If you use multiple of them, then they will be deployed as a whole. Below is the list of all Resources and Jobs that can be supplied as IN

      • version -- This is used as the seed version that you want to start from. You can also use this to reset the current version back to a specific value. This IN is optional if the job already has another source of a base version (such as a previous release job).

      • manifest/release/deploy -- You can add zero or more of these.

      • release -- You can add zero or more of these. release jobs can be INs to other release jobs. When you do this, the version produced from the incoming release job will be used as the base for the bump action instead of the IN version resource.

      • deploy -- You can add zero or more of these. Since a deploy job's output is a list of manifests, using it as an IN to a release job will result in that list of manifests being tagged with the version produced by the release job.

      • Any other Job or Resource will only participate in triggering release job but not in of the processing of it.

    • TASK -- Optional, but needs to be set to value managed if used

      • bump -- This is used to control the strategy used to bump your seed version. Subsequent executions of this Job will increment the last bumped value if it is semantically greater than the seed. If you want to reset, then just change the seed and it will start from the new seed again. Below is a list of accepted strategies and how they behave if a seed version of v1.0.0 was used.
        • major -- increments the major bit. So, first pass it will be v2.0.0 and next one will be v3.0.0.
        • minor -- increments the minor bit. So, first pass it will be v1.1.0 and next one will be v1.2.0.
        • patch -- increments the patch bit. So, first pass it will be v1.0.1 and next one will be v1.0.2.
        • alpha -- increments the alpha bit. So, first pass it will be v1.0.0-alpha and next one will be v1.0.0-alpha.1.
        • beta -- increments the beta bit. So, first pass it will be v1.0.0-beta and next one will be v1.0.0-beta.1.
        • rc -- increments the alpha bit. So, first pass it will be v1.0.0-rc and next one will be v1.0.0-rc.1.
        • final -- removes pre-release flags on your incoming version, so if your input was v4.0.0-rc.5, then output version would be v4.0.0.
  • on_start, on_success, on_failure, on_cancel, always are used to send notifications for those events. You need to provide a notification resource pointing to the provider like Slack, Email, IRC, Hipchat, etc.

A full detailed description of each tag is available on the Job Anatomy page


  • Since release jobs are managed jobs, free-form scripting is not allowed. on_start, on_success, on_failure, on_cancel and always only support NOTIFY tag.

  • A new version of the release job is created every time it is executed. Since the job executions are versioned, it is easy to Replay an old version to trigger your Assembly Line with the old settings. However, you should only do so if you understand the impact to your Assembly Lines.

Further Reading