Blueprint Settings

For an introduction to blueprints, please see the getting started guide. This reference covers the available template attributes that blueprint authors can use, along with some examples. For ease of reading, this reference for blueprint fields is broken down into a few logical sections.

The Blueprint system uses mustache templating under the hood, which is available to operators to specify the Blueprint definitions. This is convenient as it allows users to specify logic-free templates, without prescribing how or where workloads should be deployed. From Nelson’s perspective, it simply applies the template to produce a byte stream, and it sends that to the specified target scheduler, without explicitly knowing anything about the contents.

The renderable attributes typically take the following fashion:

Values

The most simplistic values to render are String and Int. In both of these cases, these simply work by specifying a reference to the field you want to render.

// render a field as a string
stackName: "{{stack_name}}"
// render a numeric field
completions: {{desired_instances}}

Enumerators

Some of the more interesting cases come in the form of an enumerator. In general, these work with a “context block”, which allows you to specify content in terms of key:[item, item].

{{#envvars}}
env:
  {{#envvars_list}}
  - name: "{{envvar_name}}"
    value: "{{envvar_value}}"
  {{/envvars_list}}
{{/envvars}}

In this example, we would say that envvars has the context *, because it can be used anywhere. Comparatively, the envvars_list can only be used in the context envvars because it is not otherwise accessible.

Units and Deployments

Plans