Your team has developed a Docker-based custom action that lints code and runs tests. The action is distributed privately within your organization via a GitHub Container Registry. During workflow runs, the action pulls the Docker image successfully on Linux runners, but fails intermittently on self-hosted Windows runners with image manifest errors. You suspect the issue is related to image registry authentication and runner configuration.
Which factors should you investigate and configure to resolve Docker image pull failures on self-hosted Windows runners?
Correct Answer:
A
Docker-based custom actions require the Docker runtime to be installed and running on the runner, authentication credentials to access the image registry, and network connectivity to pull the image. On self-hosted Windows runners, Docker Desktop or Docker Engine for Windows must be installed separately (unlike Linux runners, which typically have Docker pre-installed). Runner authentication to GHCR is typically automatic via GITHUB_TOKEN if configured, but network connectivity and firewall rules (including IP allow lists) can block pulls. Investigating these factors directly addresses the intermittent pull failures.
The other options are incorrect because: action inputs for registry credentials add complexity and don't solve infrastructure issues, Docker Hub is not more reliable than GHCR for authentication, the setup-docker action doesn't exist in the standard GitHub Actions ecosystem and caching Docker images doesn't address authentication failures, and converting to composite actions changes the action type fundamentally and is only necessary if Docker support is truly unavailable (not the case on Windows runners with Docker Engine installed). The correct answer addresses the root causes of the failure.
