pants-py-deploy 0.0.15rc3

Make it easy to maintain docker-compose files and helm-charts across projects with pants-py-deploy

Pants py_deploy plugin

• Purpose: Make it easy to maintain docker-compose files and helm-charts across projects
• Goals
  • Support updating services.{service_name}.
    • environment by scanning source code
    • ports by scanning source code
  • Support updating {chart}/
    • values.yaml
    • {deployment|statefulset}.yaml
  • Use Pantsbuild pants to write a plugin that can be pip installed
  • Then invoking it should be pants fix ::
  • Less pants boilerplate and support multi-platform (arm+amd) with a macro
  • Enabling docker-compose|helm-chart templates by using the macro and adding fields to pex_binary

Draft api

• BUILD file

py_deploy(
    name="app-name",
    entrypoint=package_name,
    docker={},
    helm={},# support more config in the future
)

• generates
  • 4 pex_binary targets
    • Each platform has one pex_binary for sources and one pex_binary for requirements
  • 2 docker_image targets
    • Both pex_binaries are copied, copies requirements 1st to take advantage of docker build cache
  • 1 helm_chart with a resource of all the files u
• Usage in pants publish ::
  • docker and helm repository = app-name
  • tags:, app-name-amd, app-name-arm, app-name-chart
• Usage in pants fix ::
  • docker-compose.yaml file & chart/* are create if they don't exist
  • if they exist: ensure environment & ports are up-to-date

Later ideas

• support passing env_config=dict(ignore_globs=[], include_globs=[])
• support analyzing env-vars and using them for adding services like postgres/rabbitmq/mongo etc. to the docker-compose file