Skip to content

Latest commit

 

History

History
163 lines (139 loc) · 5.44 KB

integration_test_charm.md

File metadata and controls

163 lines (139 loc) · 5.44 KB
Raw

Workflow file: integration_test_charm.yaml

Usage

Step 1: Create your workflow

# Copyright 2023 Canonical Ltd.
# See LICENSE file for licensing details.
jobs:
  build:
    name: Build charm
    uses: canonical/data-platform-workflows/.github/workflows/build_charm.yaml@v0.0.0
    with:
      cache: true

  integration-test:
    name: Integration test charm
    needs:
      - build
    uses: canonical/data-platform-workflows/.github/workflows/integration_test_charm.yaml@v0.0.0
    with:
      artifact-prefix: ${{ needs.build.outputs.artifact-prefix }}
      cloud: lxd
      juju-agent-version: 0.0.0

Step 2: Install plugins for pytest-operator (Poetry)

Step A

Add

pytest-operator-cache = {git = "https://github.com/canonical/data-platform-workflows", tag = "v0.0.0", subdirectory = "python/pytest_plugins/pytest_operator_cache"}
pytest-operator-groups = {git = "https://github.com/canonical/data-platform-workflows", tag = "v0.0.0", subdirectory = "python/pytest_plugins/pytest_operator_groups"}

to your integration test dependencies in pyproject.toml.

Step B

Disable Poetry's parallel installation for integration test dependencies.

Example tox.ini:

[testenv:integration]
set_env =
    {[testenv]set_env}
    # Workaround for https://github.com/python-poetry/poetry/issues/6958
    POETRY_INSTALLER_PARALLEL = false

Step C

If you're using tox, pass in the CI and GITHUB_OUTPUT environment variables in tox.ini.

[testenv:integration]
pass_env =
    {[testenv]pass_env}
    CI
    GITHUB_OUTPUT

Step 3: Split test functions into groups

Groups allow a test file (Python module) to be split across parallel GitHub runners. Each group (for each file) gets its own runner.

Add

@pytest.mark.group(1)

to every test function. Replace 1 with the group ID.

Deciding how to split tests into groups

Take a look at this Discourse post: https://discourse.charmhub.io/t/faster-ci-results-by-running-integration-tests-in-parallel/8816

(Optional) Select different runner for a test

Use a different runs-on string with

@pytest.mark.runner("ubuntu-20.04")

For self-hosted runners, use

@pytest.mark.runner(["self-hosted", "linux", "X64", "jammy", "large"])

Warning

Support for self-hosted runners is very limited.1

(Optional) Step 4: Add secrets

Step A

Pass in a string representation of a Python dict[str, str] built from multiple GitHub secrets.

Do not put the string into a single GitHub secret—build the string from multiple GitHub secrets so that GitHub is more likely to redact the secrets in GitHub Actions logs.

jobs:
  # ...
  integration-test:
    # ...
    uses: canonical/data-platform-workflows/.github/workflows/integration_test_charm.yaml@v0.0.0
    with:
      # ...
    secrets:
      # GitHub appears to redact each line of a multi-line secret
      # Avoid putting `{` or `}` on a line by itself so that it doesn't get redacted in logs
      integration-test: |
        { "AWS_ACCESS_KEY_ID": "${{ secrets.AWS_ACCESS_KEY_ID }}",
          "AWS_SECRET_ACCESS_KEY": "${{ secrets.AWS_SECRET_ACCESS_KEY }}", }

Python code to verify the string format:

import ast
secrets = ast.literal_eval("")
assert isinstance(secrets, dict)
for key, value in secrets.items():
    assert isinstance(key, str) and isinstance(value, str)

Step B (Poetry)

Add

pytest-github-secrets = {git = "https://github.com/canonical/data-platform-workflows", tag = "v0.0.0", subdirectory = "python/pytest_plugins/github_secrets"}

to your integration test dependencies in pyproject.toml.

Step C

If you're using tox, pass in the SECRETS_FROM_GITHUB environment variable in tox.ini.

[testenv:integration]
pass_env =
    {[testenv]pass_env}
    # ...
    SECRETS_FROM_GITHUB

Step D

Access the secrets from the github_secrets pytest fixture.

def test_foo(github_secrets):
    do_something(
        access_key_id=github_secrets["AWS_ACCESS_KEY_ID"],
        secret_access_key=github_secrets["AWS_SECRET_ACCESS_KEY"],
    )

(Optional) Step 5: Add microceph

Step A

Add

pytest-microceph = {git = "https://github.com/canonical/data-platform-workflows", tag = "v0.0.0", subdirectory = "python/pytest_plugins/microceph"}

to your integration test dependencies in pyproject.toml.

Step B

Access the S3 connection information from the microceph pytest fixture.

import pytest_microceph


def test_foo(microceph: pytest_microceph.ConnectionInformation):
    do_something(
        access_key_id=microceph.access_key_id,
        secret_access_key=microceph.secret_access_key,
        bucket=microceph.bucket,
    )

Footnotes

  1. Self-hosted runners are more subject to breaking changes, may lack feature support, and will not be tested for regressions. For troubleshooting with self-hosted runners, please contact IS on Mattermost before contacting the maintainers of this repository. Bug reports on GitHub-hosted runners will have higher priority—if you find a bug that is not specific to self-hosted runners, please reproduce on GitHub-hosted runners before reporting. ↩