Strategy syntax reference

Operations

inject

inject reads data specified as a string variable and adds its content to the main object.

inject:
  - $common
  - team_data

Let's assume the variables have the following values:

common:
  content:
    common_var: common_test
team_data:
  content:
    var: team1

The result:

content:
    common_var: common_test
    var: team1

The inject operation supports object key unpacking using the dot notation.

To unpack the content key from example above:

inject:
  - team_data.content

The result:

var: team1

add_parent

add_parent modifies the current object by adding a parent object.

inject:
  - $common
  - team_data
add_parent: parent_key

common:
  content:
    common_var: common_test
team_data:
  content:
    var: team1

The result:

prent_key:
    content:
        common_var: common_test
        var: team1

for_each

for_each loops are used to build new objects based on data read from variables.

Standard for_each loop:

By default this loops adds the key it iterates over to the main result.

directory_data:
  team1:
    data:
      path: team1
      organization: organization1
  team2:
    data:
      path: team2
      organization: organization2

for_each:
  iterator: $directory_data
  alias: directory_iterator
  actions:
    - inject:
          - $each_value.data

The result

team1: # <- this key was added automatically by the loop 
  path: team1
  organization: organization1
team2: # <- this key was added automatically by the loop 
  path: team2
  organization: organization2

iterator - an object the loops iterates over

alias - it provides a name for the loop, the alias can be used to access the loop's data from an inner for loop. Without aliases it would not be possible to access data in the loop above the immediate parent as the immediate parent is accessed with $each_key and $each_value. An alias is accessed by using $ in front of alias name, and each_key and each_value are available by appending to the alias name with # in between. Examples are $<alias_name>#each_key and $<alias_name>#each_value.

actions - a list of actions that will be applied within the loops. Actions: inject, add_parent, for_each

$each_value - an object in the current iteration. In the example above this is how each_value looks for the first iteration:

    path: team1
    organization: organization1

$each_key - a string in the current iteration. In the example above the first iteration each_key=path

List iterator for_each loop:

list_iterator loop iterates over string literals.

common_resources:
  resources:
    type_x:
      common_resource_x_1:
        type: resource
      common_resource_x_2:
        type: resource
    type_y:
      common_resource_y_1:
        type: resource
      common_resource_y_2:
        type: resource
team_data:
  resources:
    type_x:
        team_resource_x_1:
          type: team_resource
    type_y:
        team_resource_y_1:
          type: team_resource

for_each:
  list_iterator: [type_x, type_y]
  alias: list_iterator
  actions:
    inject:
      - $common.resources.$each_value
      - team_data.resources.$each_value

The result:

common_resource_x_1:
  type: resource
common_resource_x_2:
  type: resource
team_resource_x_1:
  type: team_resource
common_resource_y_1:
  type: resource
common_resource_y_2:
  type: resource
team_resource_y_1:
  type: team_resource

IMPORTANT: this type of for loop returns only each_value

Variables

data

data loads data from the specified path. If the path is a directory, it loads all the files in that directory and stores them as one big object. Alternatively, an individual file can be specified as below.

variable_definitions:
  common:
    path: /common/ec2.yaml
    type: data

data_with_key_from_path

data_with_key_from_path extends the data type functionality. The loaded data is organized based on a variable in curly braces. This data type is useful for for_each loops.

variable_definitions:
    common:
      path: /teams/{team}
      type: data_with_key_from_path

Let's assume that we have the following paths:

.
├── team1
│  ├── nested
│  │  └── second.yaml
│  └── one.yaml
└── team2
    ├── nested
    │  └── second.yaml
    └── one.yaml

The loaded data for that example is:

team1:
  <merged_files_for_team1>
team2:
  <merged_files_for_team2>

data_with_key_from_object_field

data_with_key_from_object_field extends the base data type functionality. Apart from loading data it allows to manipulate the loaded object, like choosing only a part of the object and adding parents to it.

variable_definitions:
    common:
      path: /common/ec2.yaml
      type: data_with_key_from_object_field
      source_object_path: ec2
      destination_object_path: my_resource

ec2.yaml

ec2:
  instance_type: t3.medium
  owner: Mike

The loaded that would be:

my_resource:
  instance_type: t3.medium
  owner: Mike

Where Next?

Learn all the CLI commands here.