Content processing CI/CD is flaky

[aj-stein-gsa]

This relates to ...

• [ ] the FedRAMP OSCAL Registry
• [ ] the FedRAMP OSCAL baselines
• [X] the Guide to OSCAL-based FedRAMP Content
• [ ] the Guide to OSCAL-based FedRAMP System Security Plans (SSP)
• [ ] the Guide to OSCAL-based FedRAMP Security Assessment Plans (SAP)
• [ ] the Guide to OSCAL-based FedRAMP Security Assessment Results (SAR)
• [ ] the Guide to OSCAL-based FedRAMP Plan of Action and Milestones (POA&M)
• [ ] the FedRAMP SSP OSCAL Template (JSON or XML Format)
• [ ] the FedRAMP SAP OSCAL Template (JSON or XML Format)
• [ ] the FedRAMP SAR OSCAL Template (JSON or XML Format)
• [ ] the FedRAMP POA&M OSCAL Template (JSON or XML Format)
• [ ] the FedRAMP OSCAL Validations

What happened?

It seems the content generation and testing CI in GHA is flaky looking at #641 failing on an unrelated change and clicking back through the master branch from time to time.

Recent GHA job output leads me to one hypothesis, that is some parts of how to run Docker Compose has changed and the old way is no longer supported.

In the last few years, Docker Compose had a single wrapper command executable, docker-compose and then moved to docker compose (observe the space). It would seem that GitHub discussion chatter confirms that was obsoleted.

This would mean local tests work, but not GHA. We need to investigate 1) if this is the only root case; 2) confirm/deny other root causes; 3) fix some/none/all and improve stability of all tests on PR and post-merge.

Relevant log output

##[debug]Evaluating condition for step: 'Build content generator'
##[debug]Evaluating: success()
##[debug]Evaluating success:
##[debug]=> true
##[debug]Result: true
##[debug]Starting: Build content generator
##[debug]Loading inputs
##[debug]Loading env
Run make init-content
##[debug]/usr/bin/bash -e /home/runner/work/_temp/4b01530c-f811-45dd-82e9-e2bd38a89b21.sh
Building Docker image for OSCAL content generation...
docker-compose -f vendor/oscal/build/docker-compose.yml -f src/content/docker-compose.yml build
make: docker-compose: Command not found
make: *** [src/content/module.mk:16: init-content] Error 127
Error: Process completed with exit code 2.
##[debug]Finishing: Build content generator

How do we replicate this issue?

1. Create a PR with any change.
2. Observe somewhat consistent failure of the content-artifacts.yml workflow.

Where, exactly?

GitHub Actions, so .github/workflows.

Other relevant details

No response

[aj-stein-gsa]

I actually "completed the work," but it is not merged up to master yet so I will leave this open and mark "in progress" until it is merged there and complete.