cloudposse/atmos 1.2.1

on GitHub

what

• Update stacks processor

why

• Correctly handle the case when the provided stack name matches not only a top-level YAML config file, but also a file in the (excluded) catalog (it was finding the matching catalog file (by the suffix) and then throwing the error that the stack does not exist). Now it correctly excludes all the paths defined in excluded_paths CLI config. In the stack config in the image below, atmos terraform plan test/test-component-override -s=tenant1-ue2-dev correctly excludes the file tenant1-ue2-dev.yaml in the catalog and only finds the file tenant1/ue2/dev.yaml as the top-level stack

• Process all imports in the order they are defined. Don't process imports concurrently in gorotines since the behaviour is not deterministic (whatever gorotine finishes last, the result of its import processing will be used as the final result in deep-merging). Using concurrent execution with gorotines makes it marginally (not noticeable) faster, but makes the final deep-merged result unpredictable. Now these imports

import:
  - globals/tenant1-globals
  - globals/ue2-globals
  - catalog/terraform/top-level-component1
  - catalog/terraform/test-component
  - catalog/terraform/test-component-override
  - catalog/terraform/vpc
  - catalog/terraform/tenant1-ue2-dev

vars:
  stage: dev

are processed in the same order as they are defined, and the final result for a variable is always taken from the last import catalog/terraform/tenant1-ue2-dev