The docker_image_availability ansible health check did not apply any intelligence for when registries are not reachable, and also did not search for images in the local index with all fully-qualified names.
This check takes a very long time to run in disconnected installs if not all of the required images are imported and tagged a certain way, causing it to consult the default registry which is not reachable and taking a long time to timeout for each image.
Update docker_image_availability check to:
1. Check correctly for image in docker index (using all registry names).
2. Inspect registries in the order configured, to enable finding required images in a local registry before consulting a public one.
3. Probe for connectivity to registries and don't continue to inspect ones that we can't reach.
4. Retry failed registry inspections to add robustness in case of transient network problems.
This check should be a lot more robust and performance in disconnected scenarios should be much improved.