DEVELOPMENT.md

Release

How to release to PyPI

Create PR with the new agent version

• Create a pull request that updates the agent's version number in the source code here and update CHANGELOG here.

• Get approval from the amazon-codeguru-profiler team.

• Wait until the Run tests workflow passes successfully.

Publish package to Test PyPI

• Create a new pre-release with the agent version and details about what is changed, in the format of a changelog as in previous releases. Make sure you check the box with "pre-release".

• This will trigger a release to the Test Registry of PyPI, and you can track it in the Actions tab.

• You can use the agent now from the PyPI Test registry as codeguru-profiler-agent.

Publish package to Live PyPI

• If you want to release to the Live registry of PyPI to be used by customers, go to the Releases tab and edit your pre-releases as prod release.

• This will trigger publishing the package to the Live Registry of PyPI, and you can track it in the Actions tab.

• You can use the agent now from the PyPI Live registry as codeguru-profiler-agent.

How to release the Lambda Layer.

The layer is used for profiling AWS lambda functions. The layer contains only our module source code as boto3 is already available in a lambda environment.

Check internal instructions for what credentials to use.

1. Checkout the last version of the main branch locally after you did the release to PyPI.

2. Run the following command in this package to publish a new version for the layer that will be available to the public immediately.

   python release_layer.py
   

3. Update the documentation with the ARN that was printed.